Tools for Analyzing Talk Part 3: Morphosyntactic Analysis

Tools for Analyzing Talk Part 3: Morphosyntactic Analysis
Tools for Analyzing Talk
Part 3: Morphosyntactic Analysis
Brian MacWhinney
Carnegie Mellon University
August 30, 2017
When citing the use of TalkBank and CHILDES facilities, please use this reference to the
last printed version:
MacWhinney, B. (2000). The CHILDES Project: Tools for Analyzing Talk. 3 rd Edition.
Mahwah, NJ: Lawrence Erlbaum Associates
This allows us to systematically track usage of the programs and data through
Part 3: Morphosyntax
Introduction ....................................................................................................................... 4
Morphosyntactic Coding ................................................................................................ 5
Running the Program Chain ...................................................................................... 14
Morphological Analysis............................................................................................... 15
One-to-one correspondence ............................................................................................. 5
Tag Groups and Word Groups.......................................................................................... 6
Words ....................................................................................................................................... 6
Part of Speech Codes ........................................................................................................... 7
Stems ......................................................................................................................................... 8
Affixes ....................................................................................................................................... 8
Clitics......................................................................................................................................... 9
Compounds .......................................................................................................................... 10
Punctuation Marks............................................................................................................ 11
Sample Morphological Tagging for English ............................................................. 11
The Design of MOR ............................................................................................................ 15
Example Analyses .............................................................................................................. 15
Exclusions in MOR ............................................................................................................. 16
Unique Options................................................................................................................... 16
Categories and Components of MOR .......................................................................... 17
MOR Part-of-Speech Categories ................................................................................... 18
MOR Grammatical Categories ....................................................................................... 21
Compounds and Complex Forms ................................................................................. 22
Errors and Replacements ............................................................................................... 23
Affixes .................................................................................................................................... 24
Control Features and Output Features ...................................................................... 24
Correcting errors .......................................................................................................... 25
Lexicon Building ................................................................................................................ 27
Disambiguator Mode ........................................................................................................ 28
A Formal Description of the Rule Files .................................................................. 29
Declarative structure ....................................................................................................... 29
Pattern-matching symbols ............................................................................................. 29
Variable notation............................................................................................................... 30
Category Information Operators ................................................................................. 30
Arules..................................................................................................................................... 31
Crules ..................................................................................................................................... 32
Building new MOR grammars ................................................................................... 35
minMOR ................................................................................................................................ 35
Adding affixes ..................................................................................................................... 35
Interactive MOR ................................................................................................................. 36
Testing ................................................................................................................................... 36
Building Arules................................................................................................................... 37
Building crules ................................................................................................................... 38
MOR for Bilingual Corpora ........................................................................................ 41
POST................................................................................................................................... 43
POSTLIST .............................................................................................................................. 44
Part 3: Morphosyntax
POSTMODRULES ................................................................................................................ 45
POSTMORTEM .................................................................................................................... 45
POSTTRAIN .......................................................................................................................... 46
POSTMOD............................................................................................................................ 49
GRASP – Syntactic Dependency Analysis .......................................................... 50
Building a training corpus ..................................................................................... 57
GRs for other languages ......................................................................................... 79
Grammatical Relations .................................................................................................... 50
Predicate-head relations ................................................................................................ 51
Argument-head relations ............................................................................................... 53
Extra-clausal elements .................................................................................................... 55
Cosmetic relations ............................................................................................................ 55
MEGRASP .............................................................................................................................. 56
OBJ and OBJ2 ....................................................................................................................... 57
3. JCT and POBJ ................................................................................................................... 58
PRED and NJCT ................................................................................................................... 59
AUX and NEG ....................................................................................................................... 61
MOD and POSS .................................................................................................................... 62
CONJ, and COORD ............................................................................................................... 62
ENUM and LP ....................................................................................................................... 63
POSTMOD ............................................................................................................................. 65
COMP, LINK .......................................................................................................................... 66
XCOMP and INF .............................................................................................................. 68
QUANT and PQ ............................................................................................................... 68
CSUBJ, COBJ, CPOBJ, CPRED ....................................................................................... 70
CJCT and XJCT ................................................................................................................. 71
CMOD and XMOD........................................................................................................... 72
BEG, BEGP, END, ENDP ................................................................................................ 73
COM and TAG .................................................................................................................. 74
SRL, APP ........................................................................................................................... 75
NAME, DATE .................................................................................................................... 76
INCROOT, OM.................................................................................................................. 77
Spanish .................................................................................................................................. 79
Chinese .................................................................................................................................. 79
Japanese ................................................................................................................................ 80
Part 3: Morphosyntax
1 Introduction
This third volume of the TalkBank manuals deals with the use of the programs that
perform automatic computation of the morphosyntactic structure of transcripts in CHAT
format. These manuals, the programs, and the TalkBank datasets can all be downloaded
freely from
The first volume of the TalkBank manual describes the CHAT transcription format.
The second volume describes the use of the CLAN data analysis programs. This third
manual describes the use of the MOR, POST, POSTMORTEM, and MEGRASP programs
to add a %mor and %gra line to CHAT transcripts. The %mor line provides a complete
part-of-speech tagging for every word indicated on the main line of the transcript. The
%gra line provides a further analysis of the grammatical dependencies between items in
the %mor line. These programs for morphosyntactic analysis are all built into CLAN.
Users who do not wish to create or process information on the %mor and %gra lines
will not need to read this current manual. However, researchers and clinicians interested
in these features will need to know the basics of the use of these programs, as described in
the next chapter. The additional sections of this manual are directed to researchers who
wish to extend or improve the coverage of MOR and GRASP grammars or who wish to
build such grammars for languages that are not yet covered.
Part 3: Morphosyntax
2 Morphosyntactic Coding
Linguists and psycholinguists rely on the analysis of morphosyntax to illuminate core
issues in learning and development. Generativist theories have emphasized issues such as:
the role of triggers in the early setting of a parameter for subject omission (Hyams &
Wexler, 1993), evidence for advanced early syntactic competence (Wexler, 1998),
evidence for early absence functional categories that attach to the IP node (Radford, 1990),
the role of optional infinitives in normal and disordered acquisition (Rice, 1997), and the
child’s ability to process syntax without any exposure to relevant data (Crain, 1991).
Generativists have sometimes been criticized for paying inadequate attention to the
empirical patterns of distribution in children’s productions. However, work by researchers
in this tradition, such as Stromswold (1994), van Kampen (1998), and Meisel (1986),
demonstrates the important role that transcript data can play in evaluating alternative
generative accounts.
Learning theorists have placed an even greater emphasis on the use of transcripts for
understanding morphosyntactic development. Neural network models have shown how
cue validities can determine the sequence of acquisition for both morphological
(MacWhinney & Leinbach, 1991; MacWhinney, Leinbach, Taraban, & McDonald, 1989;
Plunkett & Marchman, 1991) and syntactic (Elman, 1993; Mintz, Newport, & Bever, 2002;
Siskind, 1999) development. This work derives further support from a broad movement
within linguistics toward a focus on data-driven models (Bybee & Hopper, 2001) for
understanding language learning and structure. These accounts formulate accounts that
view constructions (Tomasello, 2003) and item-based patterns (MacWhinney, 1975) as the
loci for statistical learning.
The study of morphosyntax also plays an important role in the study and treatment of
language disorders, such as aphasia, specific language impairment, stuttering, and
dementia. For this work, both researchers and clinicians can benefit from methods for
achieving accurate automatic analysis of correct and incorrect uses of morphosyntactic
devices. To address these needs, the TalkBank system uses the MOR command to
automatically generate candidate morphological analyses on the %mor tier, the POST
command to disambiguate these analyses, and the MEGRASP command to compute
grammatical dependencies on the %gra tier.
2.1 One-to-one correspondence
MOR creates a %mor tier with a one-to-one correspondence between words on the
main line and words on the %mor tier. In order to achieve this one-to-one correspondence,
the following rules are observed:
1. Each word group (see below) on the %mor line is surrounded by spaces or an initial
tab to correspond to the corresponding space-delimited word group on the main line.
The correspondence matches each %mor word (morphological word) to a main line
word in a left-to-right linear order in the utterance.
2. Utterance delimiters are preserved on the %mor line to facilitate readability and
analysis. These delimiters should be the same as the ones used on the main line.
3. Along with utterance delimiters, the satellite markers of ‡ for the vocative and „ for
tag questions or dislocations are also included on the %mor line in a one-to-one
alignment format.
Part 3: Morphosyntax
4. Retracings and repetitions are excluded from this one-to-one mapping, as are
nonwords such as xxx or strings beginning with &. When word repetitions are marked
in the form word [x 3], the material in parentheses is stripped off and the word is
considered as a single form.
5. When a replacing form is indicated on the main line with the form [: text], the material
on the %mor line corresponds to the replacing material in the square brackets, not the
material that is being replaced. For example, if the main line has gonna [: going to],
the %mor line will code going to.
6. The [*] symbol that is used on the main line to indicate errors is not duplicated on the
%mor line.
2.2 Tag Groups and Word Groups
On the %mor line, alternative taggings of a given word are clustered together in tag
groups. These groups include the alternative taggings of a word that are produced by the
MOR program. Alternatives are separated by the ^ character. Here is an example of a tag
group for one of the most ambiguous words in English:
After you run the POST program on your files, all of these alternatives will be
disambiguated and each word will have only one alternative. You can also use the hand
disambiguation method built into the CLAN editor to disambiguate each tag group case by
The next level of organization for the MOR line is the word group. Word groups are
combinations marked by the preclitic delimiter $, the postclitic delimiter ~ or the compound
delimiter +. For example, the Spanish word dámelo can be represented as
This word group is a series of three words (verb~postclitic~postclitic) combined by the ~
marker. Clitics may be either preclitics or postclitics. Separable prefixes of the type found
in German or Hungarian and other discontinuous morphemes can be represented as word
groups using the preclitic delimiter $, as in this example for ausgegangen (“gone”):
Note the difference between the coding of the preclitic “aus” and the prefix “ge” in this
example. Compounds are also represented as combinations, as in this analysis of
Here, the first characters (n|) represent the part of speech of the whole compound and the
subsequent tags, after each plus sign, are for the parts of speech of the components of the
compound. Proper nouns are not treated as compounds. Therefore, they take forms with
underlines instead of pluses, such as Luke_Skywalker or New_York_City.
2.3 Words
Beneath the level of the word group is the level of the word. The structure of each
individual word is:
Part 3: Morphosyntax
=english (optional, underscore joins words)
There can be any number of prefixes, fusional suffixes, and suffixes, but there should
be only one stem. Prefixes and suffixes should be given in the order in which they occur in
the word. Since fusional suffixes are fused parts of the stem, their order is indeterminate.
The English translation of the stem is not a part of the morphology, but is included for
convenience for non-native speakers. If the English translation requires two words, these
words should be joined by an underscore as in “lose_flowers” for French défleurir.
Now let us look in greater detail at the nature of each of these types of coding. Throughout this discussion, bear in mind that all coding is done on a word-by-word basis, where
words are considered to be strings separated by spaces.
2.4 Part of Speech Codes
The morphological codes on the %mor line begin with a part-of-speech code. The basic
scheme for the part-of-speech code is:
Additional fields can be added, using the colon character as the field separator. The
subcategory fields contain information about syntactic features of the word that are not
marked overtly. For example, you may wish to code the fact that Italian “andare” is an
intransitive verb even though there is no single morpheme that signals intransitivity. You
can do this by using the part-of-speech code v:intrans, rather than by inserting a separate
In order to avoid redundancy, information that is marked by a prefix or suffix is not
incorporated into the part-of-speech code, as this information will be found to the right of
the | delimiter. These codes can be given in either uppercase, as in ADJ, or lowercase, as
in adj. In general, CHAT codes are not case-sensitive.
The particular codes given below are the ones that MOR uses for automatic morphological tagging of English. Individual researchers will need to define a system of part-ofspeech codes that correctly reflects their own research interests and theoretical commitments. Languages that are typologically quite different from English may have to use very
different part-of-speech categories. Quirk, Greenbaum, Leech, and Svartvik (1985) explain
some of the intricacies of part-of-speech coding. Their analysis should be taken as definitive for all part-of-speech coding for English.
However, for many purposes, a more
coarse-grained coding can be used.
The following set of top-level part-of-speech codes is the one used by the MOR program. Additional refinements to this system can be found by studying the organization of
the lexicon files for that program For example, in MOR, numbers are coded as types of
determiners, because this is their typical usage. The word “back” is coded as either a noun,
verb, preposition, or adjective. Further distinctions can be found by looking at the MOR
Major Parts of Speech
Part 3: Morphosyntax
Infinitive marker to
Proper Noun
Auxiliary verb, including modals
WH words
2.5 Stems
Every word on the %mor tier must include a “lemma” or stem as part of the morpheme
analysis. The stem is found on the right hand side of the | delimiter, following any preclitics or prefixes. If the transcript is in English, this can be simply the canonical form of
the word. For nouns, this is the singular. For verbs, it is the infinitive. If the transcript is in
another language, it can be the English translation. A single form should be selected for
each stem. Thus, the English indefinite article is coded as det|a with the lemma “a” whether
or not the actual form of the article is “a” or “an.”
When English is not the main language of the transcript, the transcriber must decide
whether to use English stems. Using English stems has the advantage that it makes the
corpus more available to English-reading researchers. To show how this is done, take the
German phrase “wir essen”:
wir essen.
pro|wir=we v|ess-INF=eat .
Some projects may have reasons to avoid using English stems, even as translations. In
this example, “essen” would be simply v|ess-INF. Other projects may wish to use only
English stems and no target-language stems. Sometimes there are multiple possible translations into English. For example, German “Sie”/sie” could be either “you,” “she,” or
“they.” Choosing a single English meaning for the stem helps fix the German form.
2.6 Affixes
Affixes and clitics are coded in the position in which they occur with relation to the
stem. The morphological status of the affix should be identified by the following markers
Part 3: Morphosyntax
or delimiters: - for a suffix, # for a prefix, and & for fusional or infixed morphology.
The & is used to mark affixes that are not realized in a clearly isolable phonological
shape. For example, the form “men” cannot be broken down into a part corresponding to
the stem “man” and a part corresponding to the plural marker, because one cannot say that
the vowel “e” marks the plural. For this reason, the word is coded as n|man&PL. The past
forms of irregular verbs may undergo similar ablaut processes, as in “came,” which is coded v|come&PAST, or they may undergo no phonological change at all, as in “hit”, which
is coded v|hit&PAST. Sometimes there may be several codes indicated with the & after
the stem. For example, the form “was” is coded v|be&PAST&13s. Affix and clitic codes
are based either on Latin forms for grammatical function or English words corresponding
to particular closed-class items. MOR uses the following set of affix codes for automatic
morphological tagging of English.
Inflectional Affixes for English
adjective suffix er, r
adjective suffix est, st
noun suffix ie
noun suffix s, es
noun suffix 's, '
verb suffix s, es
verb suffix ed, d
verb suffix ing
verb suffix en
Derivational Affixes for English
adjective and verb prefix un
adverbializer ly
nominalizer er
noun prefix ex
verb prefix dis
verb prefix mis
verb prefix out
verb prefix over
verb prefix pre
verb prefix pro
verb prefix re
2.7 Clitics
Clitics are marked by a tilde, as in v|parl&IMP:2S=speak~pro|DAT:MASC:SG for Italian
“parlagli” and pro|it~v|be&3s for English “it's.” Note that part of speech coding with the |
Part 3: Morphosyntax
symbol is repeated for clitics after the tilde. Both clitics and contracted elements are coded
with the tilde. The use of the tilde for contracted elements extends to forms like “sul” in
Italian, “ins” in German, or “rajta” in Hungarian in which prepositions are merged with
articles or pronouns.
Clitic Codes for English
noun phrase post-clitic 'd
noun phrase post-clitic 'll
noun phrase post-clitic 'm
noun phrase post-clitic 're
noun phrase post-clitic 's
verbal post-clitic n't
v:aux|would, v|have&PAST
v|be&1S, v:aux|be&1S
v|be&PRES, v:aux|be&PRES
v|be&3S, v:aux|be&3S
2.8 Compounds
Here are some words that we might want to treat as compounds: sweat+shirt,
tennis+court, bathing+suit, high+school, play+ground, choo+choo+train, rock+'n’+roll,
and sit+in. There are also many idiomatic phrases that could be best analyzed as linkages.
Here are some examples: a_lot_of, all_of_a_sudden, at_last, for_sure, kind_of, of_course,
once_and_for_all, once_upon_a_time, so_far, and lots_of.
On the %mor tier it is necessary to assign a part-of-speech label to each segment of the
compound. For example, the word blackboard or black+board is coded on the %mor tier
as n|+adj|black+n|board. Although the part of speech of the compound as a whole is
usually given by the part-of-speech of the final segment, forms such as make+believe
which is coded as adj|+v|make+v|believe show that this is not always true.
In order to preserve the one-to-one correspondence between words on the main line
and words on the %mor tier, words that are not marked as compounds on the main line
should not be coded as compounds on the %mor tier. For example, if the words “come
here” are used as a rote form, then they should be written as “come_here” on the main tier.
On the %mor tier this will be coded as v|come_here. It makes no sense to code this as
v|come+adv|here, because that analysis would contradict the claim that this pair functions
as a single unit. It is sometimes difficult to assign a part-of-speech code to a morpheme. In
the usual case, the part-of-speech code should be chosen from the same set of codes used
to label single words of the language. For example, some of these idiomatic phrases can be
coded as compounds on the %mor line.
Phrases Coded as Linkages
Part 3: Morphosyntax
2.9 Punctuation Marks
MOR can be configured to recognize certain punctuation marks as whole word characters.
In particular, the file punct.cut contains these entries:
{[scat end]} "end"
{[scat beg]} "beg"
{[scat cm]} "cm"
{[scat bq]} "bq"
{[scat eq]} "eq"
{[scat bq]} “bq2”
{[scat eq]} “eq2”
When the punctuation marks on the left occur in text, they are treated as separate lexical
items and are mapped to forms such as beg|beg on the %mor tier. The “end” marker is
used to mark postposed forms such as tags and sentence final particles. The “beg” marker
is used to mark preposed forms such as vocatives and communicators. The “bq” marks the
beginning of a quote and the “eq” marks the end of a quote. These special characters are
important for correctly structuring the dependency relations for the GRASP program.
2.10 Sample Morphological Tagging for English
The following table describes and illustrates a more detailed set of word class codings
for English. The %mor tier examples correspond to the labellings MOR produces for the
words in question. It is possible to augment or simplify this set, either by creating additional
word categories, or by adding additional fields to the part-of-speech label, as discussed
previously. The entries in this table and elsewhere in this manual can always be doublechecked against the current version of the MOR grammar by typing “mor +xi” to bring up
interactive MOR and then entering the word to be analyzed.
Word Classes for English
adjective, comparative
adjective, superlative
adverb, ending in ly
adverb, intensifying
adverb, post-qualifying
adverb, locative
conjunction, coord.
conjunction, subord.
determiner, singular
determiner, plural
determiner, possessive
infinitive marker
noun, common
bigger, better
biggest, best
very, rather
enough, indeed
here, then
and, or
if, although
a, the, this
these, those
my, your, her
cat, coffee
Coding of Examples
adj|big-CP, adj|good&CP
adj|big-SP, adj|good&SP
adv:int|very, adv:int|rather
adv|enough, adv|indeed
adv:loc|here, adv:tem|then
conj:coo|and, conj:coo|or
conj:sub|if, conj:sub|although
det|a, det|this
det|these, det|those
n|cat, n|coffee
Part 3: Morphosyntax
noun, plural
noun, possessive
noun, plural possessive
noun, proper
noun, proper, plural
noun, proper, possessive
noun, proper, pl. poss.
noun, adverbial
number, cardinal
number, ordinal
pronoun, personal
pronoun, reflexive
pronoun, possessive
pronoun, demonstrative
pronoun, indefinite
pronoun, indef., poss.
verb, base form
verb, 3rd singular present
verb, past tense
verb, present participle
verb, past participle
verb, modal auxiliary
home, west
all, both
I, me, we, us, he
myself, ourselves
mine, yours, his
that, this, these
everybody, nothing
half, all
walk, run
walks, runs
walked, ran
walking, running
walked, run
can, could, must
n|home, adv:loc |home
post|all, post|both
prep|in, adv:loc|in
pro|I, pro|me, pro|we, pro|us
pro:poss|mine, pro:poss:det|his
qn|half, qn|all
v|walk, v|run
v|walk-3S, v|run-3S
v|walk-PAST, v|run&PAST
part|walk-PRESP, part|run-PRESP
part|walk-PASTP, part|run&PASTP
aux|can, aux|could, aux|must
Since it is sometimes difficult to decide what part of speech a word belongs to, we
offer the following overview of the different part-of-speech labels used in the standard
English grammar.
ADJectives modify nouns, either prenominally, or predicatively. Unitary compound modifiers such as good-looking should be labeled as adjectives.
ADVerbs cover a heterogenous class of words including: manner adverbs, which generally
end in -ly; locative adverbs, which include expressions of time and place; intensifiers
that modify adjectives; and post-head modifiers, such as indeed and enough.
COmmunicators are used for interactive and communicative forms which fulfill a variety
of functions in speech and conversation. Also included in this category are words used
to express emotion, as well as imitative and onomatopeic forms, such as ah, aw,
boom, boom-boom, icky, wow, yuck, and yummy.
CONJunctions conjoin two or more words, phrases, or sentences. Examples include:
although, because, if, unless, and until.
Part 3: Morphosyntax
COORDinators include and, or, and as well as. These can combine clauses, phrases, or
DETerminers include articles, and definite and indefinite determiners. Possessive determiners such as my and your are tagged det:poss.
INFinitive is the word “to” which is tagged inf|to.
INTerjections are similar to communicators, but they typically can stand alone as
complete utterances or fragments, rather than being integrated as parts of the utterances.
They include forms such as wow, hello, good-morning, good-bye, please, thank-you.
Nouns are tagged with n for common nouns, and n:prop for proper nouns (names of people, places, fictional characters, brand-name products).
NEGative is the word “not” which is tagged neg|not.
NUMbers are labelled num for cardinal numbers. The ordinal numbers are adjectives.
Onomatopoeia are words that imitate the sounds of nature, animals, and other noises.
Particles are words that are often also prepositions, but are serving as verbal particles.
PREPositions are the heads of prepositional phrases. When a preposition is not a part of
a phrase, it should be coded as a particle or an adverb.
PROnouns include a variety of structures, such as reflexives, possessives, personal
pronouns, deictic pronouns, etc.
QUANTifiers include each, every, all, some, and similar items.
Verbs can be either main verbs, copulas, or auxililaries.
Part 3: Morphosyntax
3 Running the Program Chain
It is possible to construct a complete automatic morphosyntacgtic analysis of a series
of CHAT transcripts through a single command in CLAN, once you have the needed
programs in the correct configuration. This command runs the MOR, POST,
POSTMORTEM, and MEGRASP commands in an automatic sequence or chain. To do
this, you follow these steps:
1. Place all the files you wish to analyze into a single folder.
2. Start the CLAN program (see the Part 2 of the manual for instructions on installing
3. In CLAN’s Commands window, click on the buttom labelled Working to set your
working directory to the folder that has the files to be analyzed.
4. Under the File menu at the top of the screen, select Get MOR Grammar and select
the language you want to analyze. To do this, you must be connected to the Internet.
If you have already done this once, you do not need to do it again. By default, the
MOR grammar you have chosen will download to your desktop.
5. If you choose to move your MOR grammar to another location, you will need to use
the Mor Lib button in the Commands window to tell CLAN about where to locate it.
6. To analyze all the files in your Working directory folder, enter this command in the
Comands window: mor *.cha
7. CLAN will then run these programs in sequence: MOR, POST, POSMORTEM, and
MEGRASP. These programs will add %mor and %gra lines to your files.
8. If this command ends with a message saying that some words were not recognized,
you will probably want to fix them. If you do not, some of the entries on the %mor
line will be incomplete and the relations on the %gra line will be less accurate. If you
have doubts about the spellings of certain words, you can look in the 0allwords.cdc
file this is included in the /lex folder for each language. The words there are listed in
alphabetical order.
9. To correct errors, you can run this command: mor +xb *.cha.. Guidelines for fixing
errors are given in chapter 4 below.
Part 3: Morphosyntax
4 Morphological Analysis
4.1 The Design of MOR
The computational design of MOR was guided by Roland Hausser’s (1990) MORPH
system and was implemented by Mitzi Morris. Since 2000, Leonid Spektor has extended
MOR in many ways. Christophe Parisse built POST and POSTTRAIN (Parisse & Le
Normand, 2000). Kenji Sagae built MEGRASP as a part of his dissertation work for the
Language Technologies Institute at Carnegie Mellon University (Sagae, MacWhinney, &
Lavie, 2004a, 2004b). Leonid Spektor then integrated the program into CLAN.
The system has been designed to maximize portability across languages, extendability
of the lexicon and grammar, and compatibility with the CLAN programs. The basic engine
of the parser is language independent. Language-specific information is stored in separate
data files that can be modified by the user. The lexical entries are also kept in ASCII files
and there are several techniques for improving the match of the lexicon to a corpus. To
maximize the complete analysis of regular formations, only stems are stored in the lexicon
and inflected forms appropriate for each stem are compiled at run time.
4.2 Example Analyses
To give an example of the results of a MOR analysis for English, consider this sentence
from eve15.cha in Roger Brown’s corpus for Eve.
oops I spilled it.
co|oops pro:subj|I v|spill-PAST pro:per|it.
Here, the main line gives the child’s production and the %mor line gives the part of speech
for each word, along with the morphological analysis of affixes, such as the past tense mark
(-PAST) on the verb. The %mor lines in these files were not created by hand. To produce
them, we ran the MOR command, using the MOR grammar for English, which can be
downloaded using the Get MOR Grammar function described in the previous chapter.
The command for running MOR by itself without running the rest of the chain is: mor +d
*.cha. After running MOR, the file looks like this:
*CHI: oops I spilled it .
%mor: co|oops pro:subj|I part|spill-PASTP^v|spill-PAST pro:per|it .
In the %mor tier, words are labeled by their syntactic category or “scat”, followed by the
pipe separator |, followed then by the stem and affixes. Notice that the word “spilled” is
initially ambiguous between the past tense and participle readings. The two ambiguities are
separated by the ^ character. To resolve such ambiguities, we run a program called POST.
The command is just “post *.cha” After POST has been run, the %mor line will only have
Using this disambiguated form, we can then run the MEGRASP program to create the
representation given in the %gra line below:
oops I spilled it .
%mor: co|oops pro:subj|I v|spill-PAST pro:per|it .
%gra: 1|3|COM 2|3|SUBJ 3|0|ROOT 4|3|OBJ 5|3|PUNCT
In the %gra line, we see that the second word “I” is related to the verb (“spilled”) through
Part 3: Morphosyntax
the grammatical relation (GR) of Subject. The fourth word “it” is related to the verb
through the grammatical relation of Object. The verb is the Root and it is related to the
“left wall” or item 0.
4.3 Exclusions in MOR
Because MOR focuses on the analysis of the target utterance, it excludes a variety of nonwords, retraces, and special symbols. Specifically, MOR excludes:
1. Items that start with &
2. Pauses such as (.)
3. Unknown forms marked as xxx, yyy, or www
4. Data associated with these codes: [/?], [/-], [/], [//], and [///].
4.4 Unique Options
do not run POST command automatically. POST will run automatically after MOR,
unless this switch is used or unless the folder name includes the word “train”.
+eS Show the result of the operation of the arules on either a stem S or stems in file @S.
This output will go into a file called debug.cdc in your library directory. Another
way of achieving this is to use the +d option inside “interactive MOR”
use pinyin lexicon format for Chinese
Run MOR in the interactive test mode. You type in one word at a time to the test
prompt and MOR provides the analysis on line. This facility makes the following
commands available in the CLAN Output window:
word - analyze this word
:q quit- exit program
:c print out current set of crules
:d display application of arules.
:l re-load rules and lexicon files
:h help - print this message
If you type in a word, such as “dog” or “perro,” MOR will try to analyze it and give
you its components morphemes. If you change the rules or the lexicon, use :l to
reload and retest. The :c and :d switches will send output to a file called debug.cdc
in your library directory.
Run MOR in the lexicon building mode. This mode takes a series of .cha files as
input and outputs a small lexical file with the extension .ulx with entries for all words
not recognized by MOR. This helps in the building of lexicons.
check lexicon mode, include word location in data files
check lexicon for ambiguous entries
check lexicon mode, including capitalized words
check lexicon for compound words conflicting with plain words
Part 3: Morphosyntax
+xp check lexicon mode, including words with prosodic symbols
+xy analyze words in lex files
4.5 Categories and Components of MOR
MOR breaks up words into their component parts or morphemes. In a relatively analytic
language like English, many words require no analysis at all. However, even in English, a
word like “coworkers” can be seen to contain four component morphemes, including the
prefix “co”, the stem, the agential suffix, and the plural. For this form, MOR will produce
the analysis: co#n:v|work-AGT-PL. This representation uses the symbols # and – to
separate the four different morphemes. Here, the prefix stands at the beginning of the
analysis, followed by the stem (n|work), and the two suffixes. In general, stems always
have the form of a part of speech category, such as “n” for noun, followed by the vertical
bar and then a statement of the stem’s lexical form.
To understand the functioning of the MOR grammar for English, the best place to begin is
with a tour of the files inside the ENG folder that you can download from the server. At
the top level, you will see these files:
1. ar.cut – These are the rules that generate allomorphic variants from the stems and
affixes in the lexical files.
2. cr.cut – These are the rules that specify the possible combinations of morphemes
going from left to right in a word.
3. debug.cdc – This file holds the complete trace of an analysis of a given word by MOR.
It always holds the results of the most recent analysis. It is mostly useful for people
who are developing new ar.cut or cr.cut files as a way of tracing out or debugging
problems with these rules.
4. docs – This is a folder containing a file of instructions on how to train POST and a
list of tags and categories used in the English grammar.
5. post.db – This is a file used by POST and should be left untouched.
6. ex.cut – This file includes analyses that are being “overgenerated” by MOR and
should simply be filtered out or excluded whenever they occur.
7. lex – This folder contains many files listing the stems and affixes of the language.
We will examine it in greater detail below.
8. sf.cut – This file tells MOR how to deal with words that end with certain special form
markers such as @b for babbling.
When examining these files and others, please note that the exact shapes of the files, the
word listings, and the rules will change over time. We recommend that users glance
through these various files to understand their contents.
The first action of the parser program is to load the ar.cut file. Next the program reads
in the files in your lexicon folder and uses the rules in ar.cut to build the run-time lexicon.
Once the run-time lexicon is loaded, the parser then reads in the cr.cut file. Additionally, if
the +b option is specified, the dr.cut file is also read in. Once the concatenation rules have
been loaded the program is ready to analyze input words. As a user, you do not need to
concern yourself about the run-time lexicon. Your main concern is about the entries in the
lexicon files. The rules in the ar.cut and cr.cut files are only of concern if you wish to have
a set of analyses and labelings that differs from the one given in the chapter of the CHAT
Part 3: Morphosyntax
manual on morphosyntactic coding, or if you are trying to write a new set of grammars for
some language.
4.6 MOR Part-of-Speech Categories
The final output of MOR on the %mor line uses two sets of categories: part-of-speech
(POS) names and grammatical categories. To survey the part-of-speech names for English,
we can take a look at the files contained inside the /lex folder. These files break out the
possible words of English into different files for each specific part of speech or compound
structure. Because these distinctions are so important to the correct transcription of child
language and the correct running of MOR, it is worthwhile to consider the contents of each
of these various files. As the following table shows, about half of these word types involve
different part of speech configurations within compounds. This analysis of compounds into
their part of speech components is intended to further study of the child’s learning of
compounds as well as to provide good information regarding the part of speech of the
whole. The name of the compound files indicates their composition. For example, the
name adj+n+adj.cut indicates compounds with a noun followed by an adjective (n+adj)
whose overall function is that of an adjective. This means that it is treated just as and
adjective (adj) by the MOR and GRASP programs. In English, the part of speech of a
compound is usually the same as that of the last component of the compound.
A few
additional part of speech (POS) categories are introduced by the 0affix.cut file. These
include: n-cl (noun clitic), v-cl (verb clitic), part (participle), and n:gerund (gerund).
Additional categories on the %mor line are introduced from the special form marker file
called sf.cut. The meanings of these various special form markers are given in the CHAT
manual. Finally, the punctuation codes bq, eq, end, beg, and cm are the POS codes used
for the special character marks given in the punct.cut file.
Part 3: Morphosyntax
File (.cut)
prefixes and suffixes
terms local to the UK
baby talk adjectives
baby talk doubles
irregular adjectives
ordinal numerals
predicative adjectives
combined adjectives
regular adjectives
temporal adverbs
det, art
combined adverbs
wh term
regular adverbs
Cantonese forms
rhymes, onomatopoeia
multiword phrases
regular communicators
combined conjunctions
deictic determiners
babytalk forms
noun combinations
duplicate nouns
irregular nouns
loan words
nouns with no singular
regular nouns
see expanded list below
fave, doofer, sixpence
dipsy, yumsy
nice+nice, pink+pink
better, furthest
abreast, remiss
close_by, lovey_dovey
tall, redundant
half+hearted, hot+crossed
super+duper, easy+peasy
dog+eared, stir+crazy
how_about, as_well
where, why
ajar, fast, mostly
half+off, slant+wise
half+way, off+shore
wo, wai, la
honey, dear, sir
by_jove, gee_whiz
blah, byebye, gah, no
even_though, in_case_that
and, although, because
this, that, the,
two, twelve
c_d, t_v, w_c
passie, wawa, booboo
cul_de_sac, seven_up
cow+cow, chick_chick
children, cacti, teeth
goyim, amigo, smuck
golashes, kinesics, scissors
dog, corner, window
big+shot, cutie+pie
four+by+four, dot+to+dot
Part 3: Morphosyntax
see file
inf, neg
combined prepositions
demonstrative pronouns
indefinite pronouns
personal pronouns
possessive pronouns
possessive determiners
interrogative pronouns
small forms
baby verbs
cliticized forms
verb duplications
irregular verbs
modal auxiliaries
regular verbs
omitted words
quack+duck, moo+cow
candy+bar, foot+race
children+bed, dog+fish
wee+wee, meow+meow
squirm+worm, snap+bead
chin+up, hide+out
boom, choo_choo
all, too
out_of, in_between
under, minus
this, that
everybody, few
he, himself
hers, mine
her, my
who, what
some, all, only, most
that, which
not, to, xxx, yyy
had, getting
wee, poo
gonna, looka
be, become
eat+eat, drip+drip
came, beset, slept
hafta, gotta
can, ought
run, take, remember
deep+fry, tippy+toe
bunny+hop, sleep+walk
0know, 0conj, 0n, 0is
The construction of these lexicon files involves a variety of decisions. Here are some of
the most important issues to consider.
Words may often appear in several files. For example, virtually every noun in
English can also function as a verb. However, when this function is indicated by a
suffix, as in “milking” the noun can be recognized as a verb through a process of
morphological derivation contained in a rule in the cr.cut file. In such cases, it is
not necessary to list the word as a verb. Of course, this process fails for unmarked
verbs. However, it is generally not a good idea to represent all nouns as verbs, since
Part 3: Morphosyntax
this tends to overgenerate ambiguity. Instead, it is possible to use the
POSTMORTEM program to detect cases where nouns are functioning as bare verbs.
If a word can be analyzed morphologically, it should not be given a full listing. For
example, since “coworker” can be analyzed by MOR into three morphemes as
co#n:v|work-AGT, it should not be separately listed in the n.cut file. If it is, then
POST will not be able to distinguish co#n:v|work-AGT from n|coworker.
In the zero.cut file, possible omitted words are listed without the preceding 0. For
example, there is an entry for “conj” and “the”. However, in the transcript, these
would be represented as “0conj” and “0the”.
It is always best to use spaces to break up word sequences that are just combinations
of words. For example, instead of transcribing 1964 as “nineteen+sixty+four”,
“nineteen-sixty-four”, or “nineteen_sixty_four”, it is best to transcribe simply as
“nineteen sixty four”. This principle is particularly important for Chinese, where
there is a tendency to underutilize spaces, since Chinese itself is written without
For most languages that use Roman characters, you can rely on capitalization to
force MOR to treat words as proper nouns. To understand this, take a look at the
forms in the sf.cut file at the top of the MOR directory. These various entries tell
MOR how to process forms like [email protected] for the letter “k” or John_Paul_Jones for the
famous admiral. The symbol \c indicates that a form is capitalized and the symbol
\l indicates that it is lowercase.
4.7 MOR Grammatical Categories
In addition to the various part-of-speech categories provided by the lexicon, MOR also
inserts a series of grammatical categories, based on the information about affixes in the
0affix.cut file, as well as information inserted by the a-rules and c-rules. If the category is
regularly attached, it is preceded by a dash. If it is irregular, it uses an amerpsand. For
English, the inflectional categories are:
nominal plural
past tense
present participle
past participle
first singular
third singular present
first and third
In addition to these inflectional categories, English uses these derivational morphemes:
Part 3: Morphosyntax
deverbal, denominal
In these examples, the features dn, dv, and dadj indicate derivation of the forms from nouns,
verbs, or adjectives.
Other languages use many of these same features, but with many additional ones,
particularly for highly inflecting languages. Sometimes these are lowercase and sometimes
upper. Here are some examples:
4.8 Compounds and Complex Forms
The lexical files include many special compound files such as n+n+n.cut or v+n+v.cut.
Compounds are listed in the lexical files according to both their overall part of speech (Xbar) and the parts of speech of their components. However, there are seven types of
complex word combinations that should not be treated as compounds.
1. Underscored words. The n-under.cut file includes 40 forms that resemble
compounds, but are best viewed as units with non-morphemic components. For
example, kool_aid and band_aid are not analytic combinations of morphemes,
although they clearly have two components. The same is true for hi_fi and
coca_cola. In general, MOR and CLAN pay little attention to the underscore
character, so it can be used as needed when a plus for compounding is not
appropriate. The underscore mark is particularly useful for representing the
combinations of words found in proper nouns such as John_Paul_Jones,
Columbia_University, or The_Beauty_and_the_Beast.
If these words are
capitalized, they do not need to be included in the MOR lexicon, since all
capitalized words are taken as proper nouns in English. However, these forms
cannot contain pluses, since compounds are not proper nouns. And please be
careful not to overuse this form.
2. Separate words. Many noun-noun combinations in English should just be written
out as separate words. An example would be “faucet stem assembly rubber gasket
Part 3: Morphosyntax
holder”. It is worth noting here that German treats all such forms as single words.
This means that different conventions have to be adopted for German in order to
avoid the need for exhaustive listing of the infinite number of German compound
3. Spelling sequences. Sequences of letter names such as “O-U-T” for the spelling
of “out” are transcribed with the suffix @k, as in [email protected]
4. Acronyms. Forms such as FBI are transcribed with underscores, as in F_B_I.
Presence of the initial capital letter tells MOR to treat F_B_I as a proper noun. This
same format is used for non-proper abbreviations such as c_d or d_v_d.
5. Products. Coming up with good forms for commercial products such as CocaCola is tricky. Because of the need to ban the use of the dash on the main line, we
have avoided the use of the dash in these names. They should not be treated as
compounds, as in coca+cola, and compounds cannot be capitalized, so Coca+Cola
is not possible. This leaves us with the option of either coca_cola or Coca_Cola.
The option coca_cola seems best, since this is not a proper noun.
6. Babbling and word play. In earlier versions of CHAT and MOR, transcribers
often represent sequences of babbling or word play syllables as compounds. This
was done mostly because the plus provides a nice way of separating out the separate
syllables in these productions. To make it clear that these separations are simply
marked for purposes of syllabification, we now ask transcribers to use forms such
as ba^ba^ga^[email protected] or choo^bung^choo^[email protected] to represent these patterns.
The introduction of this more precise system for transcription of complex forms opens up
additional options for programs like MLU, KWAL, FREQ, and GRASP. For MLU,
compounds will be counted as single words, unless the plus sign is added to the morpheme
delimiter set using the +b+ option switch. For GRASP, processing of compounds only
needs to look at the overall part of speech of the compound, since the internal composition
of the compound is not relevant to the syntax. Additionally, forms such as "faucet handle
valve washer assembly" do not need to be treated as compounds, since GRASP can learn
to treat sequences of nouns as complex phrases header by the final noun.
4.9 Errors and Replacements
Transcriptions on the main line have to serve two, sometimes conflicting (Edwards, 1992),
functions. On the one hand, they need to represent the form of the speech as actually
produced. On the other hand, they need to provide input that can be used for
morphosyntactic analysis. When words are pronounced in their standard form, these two
functions are in alignment. However, when words are pronounced with phonological or
morphological errors, it is important to separate out the actual production from the
morphological target. This can be done through a system for main line tagging of errors.
This system largely replaces the coding of errors on a separate %err line, although that
form is still available, if needed. The form of the newer system is illustrated here:
*CHI: him [* case] ated [: ate] [* +ed-sup] a f(l)ower and a pun [: bun].
For the first error, there is no need to provide a replacement, since MOR can process “him”
as a standard pronoun. However, since the second word is not a real word form, the
Part 3: Morphosyntax
replacement is necessary in order to tell MOR how to process the form. The third error is
just an omission of “l” from the cluster and the final error is a mispronunciation of the
initial consonant. Phonological errors are not coded here, since that level of analysis is best
conducted inside the Phon program (Rose et al., 2005).
4.10 Affixes
The inflectional and derivational affixes of English are listed in the 0affix.cut file.
1. This file begins with a list of prefixes such as “mis” and “semi” that attach either to
nouns or verbs. Each prefix also has a permission feature, such as [allow mis]. This
feature only comes into play when a noun or verb in n.cut or v.cut also has the feature
[pre no]. For example, the verb “test” has the feature [pre no] included in order to
block prefixing with “de-” to produce “detest” which is not a derivational form of
"test". At the same time, we want to permit prefixing with “re-”, the entry for “test”
has [pre no][allow re]. Then, when the relevant rule in cr.cut sees a verb following
“re-” it checks for a match in the [allow] feature and allows the attachment in this
2. Next we see some derivational suffixes such as diminutive –ie or agential –er. Unlike
the prefixes, these suffixes often change the spelling of the stem by dropping silent e
or doubling final consonants. The ar.cut file controls this process, and the [allo x]
features listed there control the selection of the correct form of the suffix.
3. Each suffix is represented by a grammatical category in parentheses. These categories
are taken from a typologically valid list given in the CHAT Manual.
4. Each suffix specifies the grammatical category of the form that will result after its
attachment. For suffixes that change the part of speech, this is given in the scat, as in
[scat adj:n]. Prefixes do not change parts of speech, so they are simply listed as [scat
pfx] and use the [pcat x] feature to specify the shape of the forms to which they can
5. The long list of suffixes concludes with a list of cliticized auxiliaries and reduced
main verbs. These forms are represented in English as contractions. Many of these
forms are multiply ambiguous and it will be the job of POST to choose the correct
reading from among the various alternatives.
4.11 Control Features and Output Features
The lexical files include several control features that specify how stems should be
treated. One important set includes the [comp x+x] features for compounds. This feature
controls how compounds will be unpacked for formatting on the %mor line. Irregular
adjectives in adj-ir.cut have features specifying their degree as comparative or superlative.
Irregular nouns have features controlling the use of the plural. Irregular verbs have features
controlling consonant doubling [gg +] and the formation of the perfect tense. Features like
[block ed] are used to prevent reocognition of overregularized forms such as goed.
There are also a variety of features that are included in lexical entries, but not
necessarily present in the final output. For example, the feature of gender is used to
determine patterns of suffixation in Spanish, but to include this feature in the output it must
be present and not commented in the output.cut file. Other lexical features of this type
include root, ptn, num, tense, and deriv.
Part 3: Morphosyntax
5 Correcting errors
When running MOR on a new set of CHAT files, it is important to make sure that
MOR will be able to recognize all the words in these files. A first step in this process
involves running the CHECK program to see if all the words follow basic CHAT rules,
such as not including numbers or capital letters in the middle of words. There are several
common reasons for a word not being recognized:
1. It is misspelled. If you have doubts about the spellings of certain words, you can look
in the 0allwords.cdc file this is included in the /lex folder for each language. The
words there are listed in alphabetical order.
2. The word should be preceded by and ampersand & to block look up through MOR.
There are four forms using the ampersand. Nonwords just take the & alone, as in
&gaga. Incomplete words should be transcribed as &+text, as in &+sn for the
beginning of snake. Filler words should be transcribed as &-uh. Finally, sounds like
laughing can be transcribed as &=laughs, as described more extensively in the CHAT
3. The word should have been transcribed with a special form marker, as in [email protected] or
bo^[email protected] for onomatopoeia. It is impossible to list all possible onomatopoeic forms
in the MOR lexicon, so the @o marker solves this problem by telling MOR how to
treat the form. This approach will be needed for other special forms, such as babbling,
word play, and so on.
4. The word was transcribed in “eye-dialect” to represent phonological reductions.
When this is done, there are two basic ways to allow MOR to achieve correct lookup.
If the word can be transcribed with parentheses for the missing material, as in
“(be)cause”, then MOR will be happy. This method is particularly useful in Spanish
and German. Alternatively, if there is a sound substitution, then you can transcribe
using the [: text] replacement method, as in “pittie [: kittie]”.
5. You should treat the word as a proper noun by capitalizing the first letter. This
method works for many languages, but not in German where all nouns are capitalized
and not in Asian languages, since those languages do not have systems for
6. The stem is in the lexicon, but the inflected form is not recognized. In this case, it is
possible that one of the analytic rules of MOR is not working. These problems can
be reported to [email protected]
7. The stem or word is missing from MOR. In that case, you can create a file called
something like 0add.cut in the /lex folder of the MOR grammar. Once you have
accumulated a collection of such words, you can email them to [email protected] for
permanent addition to the lexicon.
Some of these forms can be corrected during the initial process of transcription by running
CHECK. However, others will not be evident until you run the MOR command with +xb
or +xl and get a list of unrecognized words.
To correct these problems, there are basically two possible tools. The first is the
KWAL program built in to CLAN. Let us say that your filename.ulx.cex list of
unrecognized words has the form “cuaght” as a misspelling of “caught.” Let us further
imagine that you have a single collection of 80 files in one folder. To correct this error,
just type this command into the Commands window:
kwal *.cha +scuaght
Part 3: Morphosyntax
KWAL will then send input to your screen as it goes through the 80 files. There may
be no more than one case of this misspelling in the whole collection. You will see this as
the output scrolls by. If necessary, just scroll back in the CLAN Output window to find
the error and then triple click to go to the spot of the error and then retype the word
For errors that are not too frequent, this method works fairly well. However, if you
have made some error consistently and frequently, you may need stronger methods.
Perhaps you transcribed “byebye” as “bye+bye” as many as 60 times. In this case, you
could use the CHSTRING program to fix this, but a better method would involve the use
of a Programmer’s Editor system such as BBEdit for the Mac or Epsilon for Windows.
Any system you use must include an ability to process Regular Expressions (RegExp) and
to operate smoothly across whole directories at a time. However, let me give a word of
warning about the use of more powerful editors. When using these systems, particularly
at first, you may make some mistakes. Always make sure that you keep a backup copy of
your entire folder before each major replacement command that you issue.
Once you know that a corpus passes CHECK, you will want to see whether it contains
words that are either misspelled or not yet in the MOR lexicon. You do this by running
the command:
mor +xb *.cha
The output from this command will have the extension .ulx.cex. After running the
command, its name will appear at the end of the output in the CLAN Output window. If
that window tells you that “all words were found in the lexicon”, then you can proceed
with running
mor *.cha
However, if not all words are recognized, you can triple-click on the line listing ther
“Output File” and it will open the list of words not yet recognized by MOR. In any large
corpus, is extremely unlikely that every word would be listed in even the largest MOR
lexicon. Therefore, users of MOR need to understand how to supplement the basic lexicons
with additional entries. Before we look at the process of adding new words to the lexicon,
we first need to examine the way in which entries in the disk lexicon are structured.
The disk lexicon contains irregular forms of a word as well as the stems of regular
forms. For example, the verb “go” is stored in the disk lexicon, along with the past tense
“went,” since this latter form is suppletive and does not undergo regular rules. The disk
lexicon contains a series of files each with a series of lexical entries with one entry per line.
The lexicon may be annotated with comments, which will not be processed. A comment
begins with the percent sign and ends with a new line. A lexical entry consists of these
1. The surface form of the word.
2. Category information about the word, expressed as a set of feature-value pairs. Each
feature-value pair is enclosed in square brackets and the full set of feature-value pairs
is enclosed in curly braces. All entries must contain a feature-value pair that identifies
the syntactic category to which the word belongs, consisting of the feature “scat” with
an appropriate value.
3. Following the category information is information about the lemmatization of irregular forms. This information is given by having the citation form of the stem
Part 3: Morphosyntax
followed by the & symbol as the morpheme separator and then the grammatical
morphemes it contains.
4. Finally, if the grammar is for a language other than English, you can enter the English
translation of the word preceded by and followed by the = sign.
The following are examples of lexical entries:
{[scat v:aux]}
{[scat det]}
{[scat det]} "a"
{[scat v] [ir +]}
{[scat v] [tense past]}
When adding new entries to the lexicon it is usually sufficient to enter the citation form of
the word, along with the syntactic category information, as in the illustration for the word
“a” in the preceding examples. When working with languages other than English, you may
wish to add English glosses and even special character sets to the lexicon. For example, in
Cantonese, you could have this entry:
{[scat n]} =apple=
To illustrate this, here is an example of the MOR output for an utterance from Cantonese:
sik6 ping4gwo2 caang2 hoeng1ziu1 .
v|sik6=eat n|ping4gwo2=apple
n|caang2=orange n|hoeng1ziu1=banana .
In languages that use both Roman and non-Roman scripts, such as Chinese, you may also
want to add non-Roman characters after the English gloss. This can be done using this
form in which the $ sign separates the English gloss from the representation in characters.
{[scat x]} “lemmatization” =gloss$characters=
MOR will take the forms indicated by the lemmatization, the gloss, and the characters and
append them after the category representation in the output. The gloss should not contain
spaces or the morpheme delimiters +, -, and #. Instead of spaces or the + sign, you can
use the underscore character to represent compounds.
5.1 Lexicon Building
When running the mor +xb command, you may wish to run the command in the form
mor +xl. The +xb form lists each separate token of an unrecognized word, whereas the +xl
form combines all the tokens into a single type. The advantage of the +xb format is that
you can click on each occurrence and change it. However, for very common errors, the
+xl format is useful because it will allow you to see what forms should be changed globally
using the CHSTRING command.
When working with the output of ther +xb form, you must then go through this file
and determine whether to discard, complete, or modify each missing case. For example, it
may be impossible to decide what category “ta” belongs to without examining where it
occurs in the corpus. In this example, a scan of the Sarah files in the Brown corpus (from
which these examples were taken), reveals that “ta” is a variant of the infinitive marker
Part 3: Morphosyntax
yeah # (be)cause if it's gon (t)a be a [email protected] it's
got ta go that way.
This missing form can be repaired by joining got and ta into gotta, because that form is
listed in the lexicon. Alternatively, the sequence can be coded as here:
yeah # (be)cause if it's gon (t)a be a [email protected] it's
gotta [: got to] go that way.
Another common source of error is misspelling. This can be repaired by correcting the
In many other cases, you will find that some words are just missing from the lexicon.
For these, you can create a file with a name like 0morewords.cut which you add to the files
in /lex. After doing this, please send the contents of this file to [email protected], so that I
can add these missing words to the authoritative version of the lexicon.
5.2 Disambiguator Mode
When POST works smoothly, there is littlel need for hand disambiguation. However,
ambiguities within a given part of speech cannot be resolved by POST and must be
disambiguated by hand using Disambiguator Mode. Also, when developing POST for a
new language, you may find this tool useful. Toggling the Disambiguator Mode option in
the Mode menu allows you to go back and forth between Disambiguator Mode and
standard Editor Mode. In Disambiguator Mode, you will see each ambiguous interpretation
on a %mor line broken into its alternative possibilities at the bottom of the editor screen.
The user double-clicks on the correct option and it is inserted. An ambiguous entry is
defined as any entry that has the ^ symbol in it. For example, the form N|back^Prep|back
is ambiguously either the noun “back” or the preposition “back.”
By default, Disambiguator Mode is set to work on the %mor tier. However, you may
find it useful for other tiers as well. To change its tier setting, select the Edit menu and pull
down to Options to get the Options dialog box. Set the disambiguation tier to the tier you
want to disambiguate. To test all of this out, edit the sample.cha file, reset your default tier,
and then type Esc-2. The editor should take you to the second %spa line which has:
At the bottom of the screen, you will have a choice of two options to select. Once the
correct one is highlighted, you hit a carriage return and the correct alternative will be
inserted. If you find it impossible to decide between alternative tags, you can select the
UND or undecided tag, which will produce a form such as “und|drink” for the word drink,
when you are not sure whether it is a noun or a verb.
Part 3: Morphosyntax
6 A Formal Description of the Rule Files
Users working with languages for which grammar files have already been built do not
need to concern themselves with this section or the next. However, users who need to
develop grammars for new languages or who find they need to modify grammars for
existing ones will need to understand how to create the two basic rule files themselves.
You do not need to create a new version of the sf.cut file for special form markers. You
just copy this file from the English MOR grammar.
To build new versions of the arules and crules files for your language, you will need
to study the English files or files for a related language. For example, when you are
building a grammar for Portuguese, it would be helpful to study the grammar that has already been constructed for Spanish. This section will help you understand the basic principles underlying the construction of the arules and crules.
6.1 Declarative structure
Both arules and crules are written using a simple declarative notation. The following
formatting conventions are used throughout:
1. Statements are one per line. Statements can be broken across lines by placing the
continuation character \ at the end of the line.
2. Comments begin with a % character and are terminated by the new line. Comments
may be placed after a statement on the same line, or they may be placed on a separate
3. Names are composed of alphanumeric symbols, plus these characters:
Both arule and crule files contain a series of rules. Rules contain one or more clauses, each
of which is composed of a series of condition statements, followed by a series of action
statements. For a clause to apply, the input(s) must satisfy all condition statements. The
output is derived from the input via the sequential application of all the action statements.
Both condition and action statements take the form of equations. The left hand side of
the equation is a keyword, which identifies the part of the input or output being processed.
The right-hand side of the rule describes either the surface patterns to be matched or generated, or the category information that must be checked or manipulated.
The analyzer manipulates two different kinds of information: information about the
surface shape of a word, and information about its category. All statements that match or
manipulate category information must make explicit reference to a feature or features.
Similarly, it is possible for a rule to contain a literal specification of the shape of a stem or
affix. In addition, it is possible to use a pattern matching language to give a more general
description of the shape of a string.
6.2 Pattern-matching symbols
The specification of orthographic patterns relies on a set of symbols derived from the
regular expression (regexp) system in Unix. The rules of this system are:
1. The metacharacters are: * [ ] | . ! All other characters are interpreted literally.
2. A pattern that contains no metacharacters will only match itself, for example the
pattern “abc” will match only the string “abc”.
3. The period matches any character.
Part 3: Morphosyntax
4. The asterisk * allows any number of matches (including 0) on the preceding character.
For example, the pattern '.*' will match a string consisting of any number of
5. The brackets [ ] are used to indicate choice from among a set of characters. The pattern
[ab] will match either a or b.
6. A pattern may consist of a disjunctive choice between two patterns, by use of the |
symbol. For example, the pattern will match all strings which end in x, s, sh, or ch.
7. It is possible to check that some input does not match a pattern by prefacing the entire
pattern with the negation operator !.
6.3 Variable notation
A variable is used to name a regular expression and to record patterns that match it. A
variable must first be declared in a special variable declaration statement. Variable declaration statements have the format: “VARNAME = regular-expression” where VARNAME
is at most eight characters long. If the variable name is more than one character, this name
should be enclosed in parenthesis when the variable is invoked. Variables are particularly
important for the arules in the ar.cut file. In these rules, the negation operator is the up
arrow ^, not the exclamation mark. Variables may be declared through combinations of
two types of disjunction markers, as in this example for the definition of a consonant cluster
in the English ar.cut file:
O = [^aeiou]|[^aeiou][^aeiou]|[^aeiou][^aeiou][^aeiou]|qu|sq
Here, the square brackets contain the definition of a consonant as not a vowel and the bar
or turnstile symbols separate alternative sequences of one, two, or three consonants. Then,
for good measure, the patterns “qu” and “squ” are also listed as consonantal onsets. For
languages that use combining diacritics and other complex symbols, it is best to use the
turnstile notation, since the square bracket notation assumes single characters. In these
strings, it is important not to include any spaces or tabs, since the presence of a space will
signal the end of the variable.
Once declared, the variable can be invoked in a rule by using the operator $. If the
variable name is longer than a single character, the variable name should be enclosed in
parentheses when invoked. For example, the statement X = .* declares and initializes a
variable named “X.” The name X is entered in a special variable table, along with the
regular expression it stands for. Note that variables may not contain other variables.
The variable table also keeps track of the most recent string that matched a named
pattern. For example, if the variable X is declared as above, then the pattern $Xle will
match all strings that end in le. For example, the string able will match this pattern, because
ab will match the pattern named by X and le will match the literal string le. Because the
string ab is matched against the named pattern X, it will be stored in the variable table as
the most recent instantiation of X, until another string matches X.
6.4 Category Information Operators
The following operators are used to manipulate category information: ADD [feature
value], and DEL [feature value]. These are used in the category action statements. For
example, the crule statement “RESULTCAT = ADD [num pl]” adds the feature value pair
Part 3: Morphosyntax
[num pl] to the result of the concatenation of two morphemes.
6.5 Arules
The function of the arules in the arules.cut file and the additional files in the /ar folder
is to expand the entries in the disk lexicon into a larger number of entries in the on-line
lexicon. Words that undergo regular phonological or orthographic changes when combined
with an affix only need to have one disk lexicon entry. The arules are used to create online lexicon entries for all inflectional variants. These variants are called allos. For
example, the final consonant of the verb “stop” is doubled before a vowel-initial suffix,
such as “-ing.” The disk lexicon contains an entry for “stop,” whereas the online lexicon
contains two entries: one for the form “stop” and one for the form “stopp”.
An arule consists of a header statement, which contains the rulename, followed by one
or more condition-action clauses. Each clause has a series of zero or more conditions on
the input, and one or more sets of actions. Here is an example of a typical condition-action
clause from the larger n-allo rule in the English ar.cut file:
LEXCAT = [scat n]
This is a single condition-action clause, labeled by the header statement “LEX-ENTRY:”
Conditions begin with one of these two keywords:
1. LEXSURF matches the surface form of the word in the lexical entry to an abstract
pattern. In this case, the variable declaration is
Y = .*[^aeiou]
Given this variable statement, the statement “LEXSURF = $Yy” will match all lexical
entry surfaces that have a final y preceded by a nonvowel.
2. LEXCAT checks the category information given in the matched lexical item against
a given series of feature value pairs, each enclosed in square brackets and separated
by commas. In this case, the rule is meant to apply only to nouns, so the category
information must be [scat n]. It is possible to check that a feature-value pair is not
present by prefacing the feature-value pair with the negation operator !.
Variable declarations should be made at the beginning of the rule, before any of the
condition-action clauses. Variables apply to all following condition-action clauses inside a
rule, but should be redefined for each rule.
After the condition statements come one or more action statements with the label ALLO: In most cases, one of the action statements is used to create an allomorph and the other
is used to enter the original lexical entry into the run-time lexicon. Action clauses begin
with one of these three keywords:
Part 3: Morphosyntax
1. ALLOSURF is used to produce an output surface. An output is a form that will be a
part of the run-time lexicon used in the analysis. In the first action clause, a lexical
entry surface form like “pony” is converted to “ponie” to serve as the stem of the
plural. In the second action clause, the original form “pony” is kept because the form
“ALLOSURF = LEXSURF” causes the surface form of the lexical entry to be copied
over to the surface form of the allo.
2. ALLOCAT determines the category of the output allos. The statement “ALLOCAT
= LEXCAT” causes all category information from the lexical entry to be copied over
to the allo entry. In addition, these two actions add the morphological classes such as
[allo nYa] or [allo nYb] in order to keep track of the nature of these allomorphs during
the application of the crules.
3. ALLOSTEM is used to produce an output stem. This action is not necessary in this
example, because this rule is fully regular and produces a noninflected stem.
However, the arule that converts “postman” into “postmen” uses this ALLOSTEM
The result of this action is the form postman&PL that is placed into the %mor line
without the involvement of any of the concatenation rules.
There are two special category feature types that operate to dump the contents of the
arules and the lexicon into the output. These are “gen” and “proc”. The gen feature
introduces its value as a component of the stem. Thus, the entry [gen m] for the Spanish
word “hombre” will end up producing n|hombre&m. The entry [proc dim] for Chinese
reduplicative verbs wil end up producing v|kan4-DIM for the reduplicated form kan4kan4.
These methods allow allorules to directly influence the output of MOR.
Every set of action statements leads to the generation of an additional allomorph for
the online lexicon. Thus, if an arule clause contains several sets of action statements, each
labeled by the header ALLO:, then that arule, when applied to one entry from the disk
lexicon, will result in several entries in the online lexicon. To create the online lexicon, the
arules are applied to the entries in the disk lexicon. Each entry is matched against the arules
in the order in which they occur in the arules file. This ordering of arules is an extremely
important feature. It means that you need to order specific cases before general cases to
avoid having the general case preempt the specific case.
As soon as the input matches all conditions in the condition section of a clause, the
actions are applied to that input to generate one or more allos, which are loaded into the
on-line lexicon. No further rules are applied to that input, and the next entry from the disk
lexicon is then read in to be processed. The complete set of arules should always end with
a default rule to copy over all remaining lexical entries that have not yet been matched by
some rule. This default rule must have this shape:
% default rule- copy input to output
RULENAME: default
6.6 Crules
The purpose of the crules in the crules.cut file is to allow stems to combine with affixes.
Part 3: Morphosyntax
In these rules, sets of conditions and actions are grouped together into if then clauses. This
allows a rule to apply to a disjunctive set of inputs. As soon as all the conditions in a clause
are met, the actions are carried out. If these are carried out successfully the rule is
considered to have “fired,” and no further clauses in that rule will be tried.
There are two inputs to a crule: the part of the word identified thus far, called the
“start,” and the next morpheme identified, called the “next.” The best way to think of this
is in terms of a bouncing ball that moves through the word, moving items from the not-yetprocessed chunk on the right over to the already processed chunk on the left. The output of
a crule is called the “result.” The following is the list of the keywords used in the crules:
check surface of start input against some pattern
check start category information
check surface of next input against some pattern
check next category information
check that start and next match for a feature-value pair type
output category information
Here is an example of a piece of a rule that uses most of these keywords:
S = .*[sc]h|.*[zxs] % strings that end in affricates
O = .*[^aeiou]o % things that end in o
% clause 1 - special case for "es" suffix
NEXTSURF = es|-es
NEXTCAT = [scat vsfx]
This rule is used to analyze verbs that end in -es. There are four conditions that must be
matched in this rule:
1. The STARTSURF is a stem that is specified in the declaration to end in an affricate.
The STARTCAT is not defined.
2. The NEXTSURF is the -es suffix that is attached to that stem.
3. The NEXTCAT is the category of the suffix, which is “vsfx” or verbal suffix.
4. The MATCHCAT [allo] statement checks that both the start and next inputs have the
same value for the feature allo. If there are multiple [allo] entries, all must match.
The shape of the result surface is simply the concatenation of the start and next
surfaces. Hence, it is not necessary to specify this via the crules. The category information
of the result is specified via the RESULTCAT statement. The statement “RESULTCAT =
STARTCAT” causes all category information from the start input to be copied over to the
result. The statement “NEXTCAT [tense]” copies the tense value from the NEXT to the
RESULT and the statement “DEL [allo]” deletes all the values for the category [allo].
In addition to the condition-action statements, crules include two other statements: the
CTYPE statement, and the RULEPACKAGES statement. The CTYPE statement identifies
the kind of concatenation expected and the way in which this concatenation is to be marked.
This statement follows the RULENAME header. There are two special CTYPE makers:
START and END. “CTYPE: START” is used for those rules that execute as soon as one
Part 3: Morphosyntax
morpheme has been found. “CTYPE: END” is used for those rules that execute when the
end of the input has been reached. Otherwise, the CYTPE marker is used to indicate which
concatenation symbol is used when concatenating the morphemes together into a parse for
a word. The # is used between a prefix and a stem, - is used between a stem and suffix, and
~ is used between a clitic and a stem. In most cases, rules that specify possible suffixes
will start with CTYPE: -. These rules insert a suffix after the stem.
Rules with CTYPE START are applied as soon as a morpheme has been recognized.
In this case, the beginning of the word is considered as the start input, and the next input is
the morpheme first recognized. As the start input has no surface and no category
information associated with it, conditions and actions are stated only on the next input.
Rules with CTYPE END are invoked when the end of a word is reached, and they are
used to rule out spurious parses. For the endrules, the start input is the entire word that has
just been parsed, and there is no next input. Thus, conditions and actions are only stated on
the start input.
The RULEPACKAGES statement identifies which rules may be applied to the result
of a rule, when that result is the input to another rule. The RULEPACKAGES statement
follows the action statements in a clause. There is a RULEPACKAGES statement associated with each clause. The rules named in a RULEPACKAGES statement are not tried
until after another morpheme has been found. For example, in parsing the input “walking”,
the parser first finds the morpheme “walk,” and at that point applies the startrules. Of these
startrules, the rule for verbs will be fired. This rule includes a RULEPACKAGES statement
specifying that the rule which handles verb conjugation may later be fired. When the parser
has further identified the morpheme “ing,” the verb conjugation rule will apply, where
“walk” is the start input, and “ing” is the next input.
Note that, unlike the arules which are strictly ordered from top to bottom of the file,
the crules have an order of application that is determined by their CTYPE and the way in
which the RULEPACKAGES statement channels words from one rule to the next.
Part 3: Morphosyntax
7 Building new MOR grammars
7.1 minMOR
The simplest possible form of a MOR grammar is represented in the “min” grammar
that you can download from the MOR grammars page at
You can begin your work using with the sample minimal MOR grammars available
from the net. This grammar includes
the sf.cut file that all of the MOR grammars use,
a sample.cha file with a few words
a basically blank ar.cut file, because no allomorphy is yet involved,
a cr.cut file that recognizes the parts of speech you will create, along with
one rule for making plural nouns, and
a lex folder with examples of verbs, nouns, a determiner, and a suffix.
You can adjust this format to your language and use a different sample.cha file to test out
the operation of this minimal MOR grammar. This system should allow you to build up a
lexicon of uninflected stems. Try to build up separate files for each of the parts of speech
in your language.
You can test out your grammar either by running mor +d sample.cha or else using
interactive MOR with mor +xi. For example, if you use interactive MOR and type “dogs”
you should get
Result: n|dog-PL
7.2 Adding affixes
At some point, you will realize that it would be more efficient to create a system for
lexical analysis, rather than relying only on full forms. This will require you to build up a
morphological grammar. When building a morphology for a new language, it is best to
begin with a paper-and-pencil analysis of the system in which you lay out the various
affixes of the language, the classes of stem allomorphy variations, and the forces that
condition the choices between allomorphs. This work should be guided by a good
descriptive grammar of the morphology of the language. For example, we have used the
Berlitz conjugation books for French, German, Italian, and Spanish.
Once this basic groundwork is finished, you may want to focus on one part-of-speech
at a time. For example, you could begin with the adverbs, since they are often
monomorphemic. Then you could move on to the nouns. The verbs should probably come
last. You can copy the sf.cut file from English and rename it.
As you start to feel comfortable with this, you should begin to add affixes. To do this,
you need to create a lexicon file, such as aff.cut. Using the technique found in unification
grammars, you want to set up categories and allos for these affixes that will allow them to
match up with the right stems when the crules fire. For example, you might want to call
the plural a [scat nsfx] in order to emphasize the fact that it should attach to nouns. And
you could give the designation [allo mdim] to the masculine diminutive suffix -ito in
Spanish in order to make sure that it only attaches to masculine stems and produces a
masculine output.
Part 3: Morphosyntax
7.3 Interactive MOR
Once you have a simple lexicon and a set of rule files, you will begin a long process
of working with interactive MOR. When using MOR in the +xi or interactive mode, there
are several additional options that become available in the CLAN Output window. They
word - analyze this word
:q quit- exit program
:c print out current set of crules
:d display application of a rules.
:l re-load rules and lexicon files
:h help - print this message
If you type in a word, such as “dog” or “perro,” MOR will try to analyze it and give you
its component morphemes. If all is well, you can move on the next word. If it is not, you
need to change your rules or the lexicon. You can stay within CLAN and just open these
using the Editor. After you save your changes, use :l to reload and retest.
7.4 Testing
As you begin to elaborate your grammar, you will want to start to work with sets of
files. These can be real data files or else files full of test words. These files provide what
computer scientists call “regression testing”. When you shift to working with files, you will
be combining the use of interactive MOR and the +xi switch with use of the lexicon testing
facility that uses +xl and +xb. As you move through this work, make copies of your MOR
grammar files and lexicon frequently, because you will sometimes find that you have made
a change that makes everything break and you will need to go back to an earlier stage to
figure out what you need to fix. We also recommend using a fast machine with lots of
memory. You will find that you are frequently reloading the grammar using the :l function,
and having a fast machine will speed this process.
As you progress with your work, continually check each new rule change by entering
:l (colon followed by “l” for load) into the CLAN Output window. If you have changed
something in a way that produces a syntactic violation, you will learn this immediately and
be able to change it back. If you find that a method fails, you may need to rethink your
logic. Consider these factors:
1. Arules are strictly ordered. Maybe you have placed a general case before a specific
2. Crules depend on direction from the RULEPACKAGES statement.
3. There must be a START and END rule for each part of speech. If you are getting too
many entries for a word, maybe you have started it twice. Alternatively, you may
have created too many allomorphs with the arules.
4. If you have a MATCHCAT allos statement, all allos must match. The operation DEL
[allo] deletes all allos and you must add back any you want to keep.
5. Make sure that you understand the use of variable notation and pattern matching
symbols for specifying the surface form in the arules.
However, sometimes it is not clear why a method is not working. In this case, you will
want to check the application of the crules using the :c option in the CLAN Output window.
You then need to trace through the firing of the rules. The most important information is
Part 3: Morphosyntax
often at the end of this output.
If the stem itself is not being recognized, you will need to also trace the operation of
the arules. To do this, you should either use the +e option in standard MOR or else the :d
option in interactive MOR. The latter is probably the most useful. To use this option, you
should create a directory called testlex with a single file with the words you are working
with. Then run:
mor +xi +ltestlex
Once this runs, type :d and then :l and the output of the arules for this test lexicon will go
to debug.cdc. Use your editor to open that file and try to trace what is happening there.
As you progress with the construction of rules and the enlargement of the lexicon, you
can tackle whole corpora. At this point you will occasionally run the +xl analysis. Then
you take the problems noted by +xl and use them as the basis for repeated testing using the
+xi switch and repeated reloading of the rules as you improve them. As you build up your
rule sets, you will want to annotate them fully using comments preceded by the % symbol.
7.5 Building Arules
In English, the main arule patterns involve consonant doubling, silent –e, changes of y to
i, and irregulars like “knives” or “leaves.” The rules use the spelling of final consonants
and vowels to predict these various allomorphic variations. Variables such as $V or $C
are set up at the beginning of the file to refer to vowels and consonants and then the rules
use these variables to describe alternative lexical patterns and the shapes of allomorphs.
For example, the rule for consonant doubling takes this shape:
LEXCAT = [scat v], ![tense OR past perf], ![gem no] % to block putting
Here, the string $O$V$C characterizes verbs like “bat” that end with vowels followed
by consonants. The first allo will produce words like “batting” or “batter” and the second
will give a stem for “bats” or “bat”. A complete list of allomorphy types for English is
given in the file engcats.cdc in the /docs folder in the MOR grammar.
When a user types the “mor” command to CLAN, the program loads up all the *.cut
files in the lexicon and then passes each lexical form past the rules of the ar.cut file. The
rules in the ar.cut file are strictly ordered. If a form matches a rule, that rule fires and the
allomorphs it produces are encoded into a lexical tree based on a “trie” structure. Then
MOR moves on to the next lexical form, without considering any additional rules. This
means that it is important to place more specific cases before more general cases in a
standard bleeding relation. There is no “feeding” relation in the ar.cut file, since each form
is shipped over to the tree structure after matching.
Part 3: Morphosyntax
7.6 Building crules
The other “core” file in a MOR grammar is the cr.cut file that contains the rules that
specify pathways through possible words. The basic idea of crules or concatenation or
continuation rules is taken from Hausser’s (1999) left-associative grammar which specifies
the shape of possible “continuations” as a parser moves from left to right through a word.
Unlike the rules of the ar.cut file, the rules in the cr.cut file are not ordered. Instead, they
work through a “feeding” relation. MOR goes through a candidate word from left to right
to match up the current sequence with forms in the lexical trie structure. When a match is
made, the categories of the current form become a part of the STARTCAT. If the
STARTCAT matches up with the STARTCAT of one of the rules in cr.cut, as well as
satisfying some additional matching conditions specified in the rule, then that rule fires.
The result of this firing is to change the shape of the STARTCAT and to then thread
processing into some additional rules.
For example, let us consider the processing of the verb “reconsidering.” Here, the first
rule to fire is the specific-vpfx-start rule which matches the fact that “re-” has the feature
[scat pfx] and [pcat v]. This initial recognition of the prefix then threads into the specificvpfx-verb rule that requires the next item have the feature [scat v]. This rule has the feature
CTYPE # which serves to introduce the # sign into the final tagging to produce
re#part|consider-PRESP. After the verb “consider” is accepted, the RULEPACKAGE tells
MOR to move on to three other rules: v-conj, n:v-deriv, and adj:v-deriv. Each of these
rules can be viewed as a separate thread out of the specific-vpfx-verb rule. At this point in
processing the word, the remaining orthographic material is “-ing”. Looking at the
0affix.cut file, we see that “ing” has three entries: [scat part], [scat v:n], and [scat
n:gerund]. One of the pathways at this point leads through the v-conj rule. Within v-conj,
only the fourth clause fires, since that clause matches [scat part]. This clause can lead on
to three further threads, but, since there is no further orthographic material, there is no
NEXTCAT for these rules. Therefore, this thread then goes on to the end rules and outputs
the first successful parse of “reconsidering.” The second thread from the specific-vpfxverb rule leads to the n:v-deriv rule. This rule accepts the reading of “ing” as [scat n:gerund]
to produce the second reading of “reconsidering”. Finally, MOR traces the third thread
from the specific-vpfx-verb rule which leads to adj:v-deriv. This route produces no
matches, so processing terminates with this result:
Result: re#part|consider-PRESP^re#n:gerund|consider-GERUND
Later, POST will work to choose between these two possible readings of
“reconsidering” on the basis of the syntactic context. As we noted earlier, when
“reconsidering” follows an auxiliary (“is eating”) or when it functions adjectivally (“an
eating binge”), it is treated as a participle. However, when it appears as the head of an NP
(“eating is good for you”), it is treated as a gerund. Categories and processes of this type
can be modified to match up with the requirements of the GRASP program to be discussed
The process of building ar.cut and cr.cut files for a new language involves a slow
iteration of lexicon building with rule building. The problem with building up a MOR
grammar one word at a time like this is that changes that favour the analysis of one word
can break the analysis of other words. To make sure that this is not happening, it is
important to have a collection of test words that you continually monitor using mor +xl.
Part 3: Morphosyntax
One approach to this is just to have a growing set of transcripts or utterances that can be
analyzed. Another approach is to have a systematic target set configured not as sentences
but as transcripts with one word in each sentence. An example of this approach can be
found in the /verbi folder in the Italian MOR grammar. This folder has one file for each of
the 106 verbal paradigms of the Berlitz Italian Verb Handbook (2005). That handbook
gives the full paradigm of one “leading” verb for each conjugational type. We then typed
all the relevant forms into CHAT files. Then, as we built up the ar.cut file for Italian, we
designed allo types using features that matched the numbers in the Handbook. In the end,
things become a bit more complex in Spanish, Italian, and French.
1. The initial rules of the ar.cut file for these languages specify the most limited and
lexically-bound patterns by listing almost the full stem, as in $Xdice for verbs like
“dicere”, “predicere” or “benedicere” which all behave similarly, or “nuoce” which
is the only verb of its type.
2. Further in the rule list, verbs are listed through a general phonology, but often limited
to the presence of a lexical tag such as [type 16] that indicates verb membership in a
conjugational class.
3. Within the rule for each verb type, the grammar specifies up to 12 stem allomorph
types. Some of these have the same surface phonology. However, to match up
properly across the paradigm, it is important to generate this full set. Once this basic
grid is determined, it is easy to add new rules for each additional conjugational type
by a process of cut-and-paste followed by local modifications.
4. Where possible, the rules are left in an order that corresponds to the order of the
conjugational numbers of the Berlitz Handbook. However, when this order interferes
with rule bleeding, it is changed.
5. Perhaps the biggest conceptual challenge is the formulation of a good set of [allo x]
tags for the paradigm. The current Italian grammar mixes together tags like [allo vv]
that are defined on phonological grounds and tags like [allo vpart] that are defined on
paradigmatic grounds. A more systematic analysis would probably use a somewhat
larger set of tags to cover all tense-aspect-mood slots and use the phonological tags
as a secondary overlay on the basic semantic tags.
6. Although verbs are the major challenge in Romance languages, it is also important to
manage verbal clitics and noun and adjectives plurals. In the end, all nouns must be
listed with gender information. Nouns that have both masculine and feminine forms
are listed with the feature [anim yes] that allows the ar.cut file to generate both sets
of allomorphs.
7. Spanish has additional complexities involving the placement of stress marks for
infinitives and imperatives with suffixed clitics, such as dámelo. Italian has additional
complications for forms such as “nello” and the various pronominal and clitic forms.
As you progress with your work, continually check each new rule change by entering :l
(colon followed by “l” for load) into the CLAN Output window and then testing some
crucial words. If you have changed something in a way that produces a syntactic violation,
you will learn this immediately and be able to change it back. If you find that a method
fails, you should first rethink your logic. Consider these factors:
1. Arules are strictly ordered. Maybe you have placed a general case before a specific
Part 3: Morphosyntax
2. Crules depend on direction from the RULEPACKAGES statement. Perhaps you are
not reaching the rule that needs to fire.
3. There has to be a START and END rule for each part of speech. If you are getting
too many entries for a word, maybe you have started it twice. Alternatively, you may
have created too many allomorphs with the arules.
4. Possibly, you form is not satisfying the requirements of the end rules. If it doesn’t
these rules will not “let it out.”
5. If you have a MATCHCAT allos statement, all allos must match. The operation DEL
[allo] deletes all allos and you must add back any you want to keep.
6. Make sure that you understand the use of variable notation and pattern matching
symbols for specifying the surface form in the arules.
However, sometimes it is not clear why a method is not working. In this case, you will
want to check the application of the crules using the :c option in the CLAN Output window.
You then need to trace through the firing of the rules. The most important information is
often at the end of this output.
If the stem itself is not being recognized, you will need to also trace the operation of
the arules. To do this, you should either use the +e option in standard MOR or else the :d
option in interactive MOR. The latter is probably the most useful. To use this option, you
should create a directory called testlex with a single file with the words you are working
with. Then run: mor +xi +ltestlex
Once this runs, type :d and then :l and the output of the arules for this test lexicon will go
to debug.cdc. Use your editor to open that file and try to trace what is happening there.
As you progress with the construction of rules and the enlargement of the lexicon, you
can tackle whole corpora. At this point you will occasionally run the +xl analysis. Then
you take the problems noted by +xl and use them as the basis for repeated testing using the
+xi switch and repeated reloading of the rules as you improve them. As you build up your
rule sets, you will want to annotate them fully using comments preceded by the % symbol.
Part 3: Morphosyntax
8 MOR for Bilingual Corpora
It is easy to use MOR and POST to process bilingual corpora. A good sample
application of this method is for the transcripts collected by Virginia Yip and Stephen
Matthews from Cantonese-English bilingual children in Hong Kong. In these corpora,
parents, caretakers, and children often switch back and forth between the two languages.
In order to tell MOR which grammar to use for which utterances, each sentence must be
clearly identified for language. It turns out that this is not too difficult to do. First, by the
nature of the goals of the study and the people conversing with the child, certain files are
typically biased toward one language or the other. In the YipMatthews corpus, English is
the default language in folders such as SophieEng or TimEng and Cantonese is the default
in folders such as SophieCan and TimCan. To mark this in the files in which Cantonese is
predominant, the @Languages tier has this form:
yue, eng
In the files in which English is predominant, on the other hand, the tier has this form:
eng, yue
The programs then assume that, by default, each word in the transcript is in the first listed
language. This default can be reversed in two ways. First, within the English files, the
precode [- yue] can be placed at the beginning of utterances that are primarily in Cantonese.
If single Cantonese words are used inside English utterances, they are marked with the
special form marker @s. If an English word appears within a Cantonese sentence marked
with the [- yue] precode, then the @s code means that the default for that sentence
(Chinese) is now reversed to the other language (English). For the files that are primarily
in Cantonese, the opposite pattern is used. In those files, English sentences are marked as
[- eng] and English words inside Cantonese are marked by @s. This form of marking
preserves readability, while still making it clear to the programs which words are in which
language. If it is important to have each word explicitly tagged for language, the –l switch
can be used with CLAN programs such as KWAL, COMBO, or FIXIT to insert this more
verbose method of language marking.
To minimize cross-language listing, it was also helpful to create easy ways of
representing words that were shared between languages. This was particularly important
for the names of family members or relation names. For example, the Cantonese form 姐
姐 for “big sister” can be written in English as Zeze, so that this form can be processed
correctly as a proper noun address term. Similarly, Cantonese has borrowed a set of
English salutations such as “byebye” and “sorry” which are simply added directly to the
Cantonese grammar in the co-eng.cut file.
Once these various adaptations and markings are completed, it is then possible to run
MOR in two passes on the corpus. By default, MOR excludes lines marked with the form
[- *] at the beginning. So, this means that, for the English corpora, the steps are:
1. Set the MOR library to English and run: mor *.cha +1
2. Disambiguate the results with: post *.cha +1
3. Run CHECK to check for problems.
4. Set the MOR library to Cantonese and run: mor +s”[- yue]” *.cha +1
5. Disambiguate the results with: post *.cha +1
6. Run CHECK to check for problems.
Part 3: Morphosyntax
To illustrate the result of this process, here is a representative snippet from the
te951130.cha file in the /TimEng folder. Note that the default language here is English and
that sentences in Cantonese are explicitly marked as [- yue].
where is grandma first, tell me ?
adv:wh|where v|be n|grandma adv|first v|tell pro|me ?
well, what's this ?
co|well pro:wh|what~v|be pro:dem|this ?
[- yue] xxx 呢 個 唔 夠 架 .
unk|xxx det|ni1=this cl|go3=cl neg|m4=not adv|gau3=enough
sfp|gaa3=sfp .
[- yue] 呢 個 唔 夠 .
det|ni1=this cl|go3=cl neg|m4=not adv|gau3=enough .
<what does it mean> [>] ?
pro:wh|what v:aux|do pro|it v|mean ?
This type of analysis is possible whenever MOR grammars exist for both languages, as
would be the case for Japanese-English, Spanish-French, Putonghua-Cantonese, or ItalianChinese bilinguals.
Part 3: Morphosyntax
POST was written by Christophe Parisse of INSERM, Paris for automatically
disambiguating the output of MOR. The POST package is composed of four CLAN
commands: POST, POSTTRAIN, POSTLIST, and POSTMOD. POST is the command
that runs the disambiguator. It uses a database called post.db that contains information
about syntactic word order. Databases are created and maintained by POSTTRAIN and can
be dumped in a text file by POSTLIST. POSTMODRULES is a utility for modifying Brill
There are POST databases now for Chinese, Japanese, Spanish, and English. As our
work with POST progresses, we will make these available for additional languages. To run
POST, you can use this command format :
post *.cha
The accuracy of disambiguation by POST for English will be between 95 and 97 percent.
This means that there will be some errors.
The options for POST are:
do not use Brill rules (they are used by default)
use a slower but more thorough version of Brill's rules analysis.
output all affixes (default)
output the affixes listed in file F and post.db. If there is a posttags.cut file, then
it is used by default as if the +cposttags.cut switch were being used.
output only the affixes defined during training with POSTTRAIN
omit the affixes in file F, but not the affixers defined during training with
use POST database file F (default is "post.db"). This file must have been created
by POSTTRAIN. If you do not use this switch, POST will try to locate a file
called post.db in either the current working directory or your MOR library
+e[1,2]c this option is a complement to the option +s2 and +s3 only. It allows you to
change the separator used (+e1c) between the different solutions, (+e2c) before
the information about the parsing process. (c can be any character). By default,
the separator for +e1 is # and for +e2, the separator is /.
send output to file derived from input file name. If you do not use this switch,
POST will create a series of output files named *.pst.
send output to file F. This switch will change the extension to the output files.
Part 3: Morphosyntax
send output to the screen
reduce memory use (but longer processing time)
when followed by a number the +l switch controls the number of output lines
tries to process unknown words.
N=0 (default) replace ambiguous %mor lines with disambiguated ones
N=1 keep ambiguous %mor lines and add disambiguated %pos lines.
N=2 output as in N=1, but with slashes marking undecidable cases.
N=3 keep ambiguous %mor lines and add %pos lines with debugging info.
N=4 inserts a %nob line before the %mor/%pos line that presents the results of
the analysis without using Brill rules.
N=5 outputs results for debuging POST grammars.
N=6 complete outputs results for debugging POST grammars.
With the options +s0 and +s1, only the best candidate is outputted. With option
+s2, second and following candidates may be outputted, when the
disambiguation process is not able to choose between different solutions with the
most probable solution displayed first. With option +s3, information about the
parsing process is given in three situations: processing of unknown words (useful
for checking these words quickly after the parsing process), processing of
unknown rules and no correct syntactic path obtained (usually corresponds to
new grammatical situations or typographic errors).
include tier code S
exclude tier code S
+/-t#Target_Child - select target child's tiers
+/[email protected]="*|Mother|*" - select mother's tiers
POSTLIST provides a list of tags used by POST. It is run on the post.db database file.
The options for POSTLIST are as follows:
this gives the name of the database to be listed (default value: ‘eng.db’).
specify name of result file to be F.
outputs all the matrix entries present in the database.
outputs all the rules present in the database.
outputs rule dictionary for the Brill tagger.
outputs rule dictionary for the Brill tagger in numerical order.
outputs the list of all tags present in the database.
outputs all the word frequencies gathered in the database.
outputs word dictionary for the Brill tagger.
If none of the options is selected, then general information about the size of the database is
Part 3: Morphosyntax
Multicat categories provide a way to pack categories together so as to create artificial
categories that have a longer context (four words instead of three) - this makes some sense
from the linguistic point of view, it is a way to consider that clitics are in fact near-flexions
and that the grammatical values is included in the word (and that in fact 'he is' is to be
considered as different from 'he's', which is oral language may not be false). However if I
had to redo all this, I would say the clitic / flexion distinction (whatever the theoretical
interpretation) should in fact be handled at MOR level, not at POST level. POST should be
get the same normalized forms, not hardcode whether they are different or dissimilar. This
would be more language independent.
The +r option outputs rules using the following conventions.
1. pct (punctuation) indicates beginning or end of the sentence.
2. multicat// indicates that the following categories are options
3. the numbers indicate the numbers of contexts for a particular member of the multicat
set, followed by the numbers of occurrences in that context
4. n:gerund|play-GERUND^part|play-PRESP^v:n|play-PRESP => 3 [0,6,0] means that
play has 3 potential categories n:gerund|play-GERUND and part|play-PRESP and
v:n|play-PRESP and that it was found 6 time in the second category in the training
This program outputs the rules used by POST for debugging rules. POSTMODRULES is
used to check and modify Brill rules after initial training.
This program relies on a dictionary file called postmortem.cut to alter the part-of-speech
tags in the %mor line after the final operation of MOR and POST. The use of this program
is restricted to cases of extreme part-of-speech extension, such as using color names as
nouns or common nouns as verbs. Here is an example of some lines in a postmortem.cut
det adj v => det n v
det adj $e => det n $e
Here, the first line will change a sequence such as “the red is” from “det adj v” to “det n
v”. The second line will change “det adj” to “det n” just in the case that the adjective is at
the end of the sentence. The symbol $e represents the end of the utterance.
The rules in POSTMORTEM vary markedly from language to language. They are
particuarly important for German, where they are used to flesh out the morphological
features of the noun phrase. In English, they are used to deal with the noun-verb ambiguity
problem. For this, the rules change the code nx to n and vx to v. The nx and vx codes are
used to make sure that a noun that almost always serves as a verb can still be “forcetranscribed” as a noun and then fixed later. To understand this, you have to look in the
Part 3: Morphosyntax
nx.cut file where you would see the entry for nxbreak. The word break is almost always a
verb, but when it is really a noun, then the transcriber can write nxbreak and after the
running of POSTMORTEM it will appear in the %mor line as n|break. The opposite is
true for the words in vx.cut, such as “vxfather” which is almost always a noun.
POSTMORTEM uses the +a switch to run in three possible modes. If no +a switch is used,
then it does all replacements automatically. Second, if you use the +a switch, then it inserts
the new string after the old one and you can then disambiguate the whole file by hand using
escape-2. Third, if you use the +a1 switch, it will work interactively and you can choose
whether to make each replacement on a case by case basis.
POSTTRAIN was written by Christophe Parisse of INSERM, Paris. In order to run
POST, you need to create a database file for your language. For several languages, this
has already been done. If there is no POST database file for your language or your subject
group, you can use the POSTTRAIN program to create this file. The default name for this
file is eng.db. If you are not working with English, you should choose some other name
for this file. Before running POSTTRAIN, you should take these steps:
1. You should specify a set of files that will be your POSTTRAIN training files. You
may wish to start with a small set of files and then build up as you go.
2. You should verify that all of your training files pass CHECK.
3. Next, you should run MOR with the +xl option to make sure that all words are
4. You then run MOR on your training files. This will produce an ambiguous %mor
5. Now you open each file in the editor and use the Esc-2 command to disambiguate the
ambiguous %mor line.
6. Once this is done for a given file, using the Query-Replace function to rename %mor
to %trn.
7. After you have created a few training files or even after you have only one file, run
MOR again.
8. Now you can run POSTTRAIN with a command like this:
posttrain +c +o0err.cut *.cha
9. Now, take a look at the 0err.cut file to see if there are problems. If not, you can test
out your POST file using POST. If the results seem pretty good, you can shift to eyebased evaluation of the disambiguated line, rather than using Esc-2. Otherwise, stick
with Esc-2 and create more training data. Whenever you are happy with a
disambiguated %mor line in a new training file, then you can go ahead and rename it
to %trn.
10. The basic idea here is to continue to improve the accuracy of the %trn line as a way
of improving the accuracy of the .db POST database file.
When developing a new POST database, you will find that eventually you need to
repeatedly cycle through a standard sets of commands while making continual changes to
the input data. Here is a sample sequence that uses the defaults in POST and POSTTRAIN:
Part 3: Morphosyntax
mor *.cha +1
posttrain +c +o0err.cut +x *.cha
post *.cha +1
trnfix *.cha
In these commands, the +1 must be used carefully, since it replaces the original. If a
program crashes or exits while running with +1, the original can be destroyed, so make a
backup of the whole directory first before running +1. TRNFIX can be used to spot
mismatches between the %trn and %mor lines.
The options for POSTTRAIN are:
train word frequencies even on utterances longer than length 3.
extended learning using Brill's rules
Brill's rules training only
+boF append output of Brill rule training to file F (default: send it to screen)
parameter for Brill rules
1- means normal Brill rules are produced (default)
2- means only lexical rules are produced
3- same as +b1, but eliminates rules redundant with binary rules
4- same as +b2, but eliminates rules redundant with binary rules
+btN threshold for Brill rules (default=2). For example, if the value is 2, a rule should
correct 3 errors to be considered useful. To generate all possible rules, use a
threshold of 0.
create new POST database file with the name post.db (default)
create new POST database file with the name F
add to an existing version of post.db
add to an existing POST database file with the name F
the affixes and stems in file F are used for training. The default name of this file
is tags.cut. So, if you want to add stems for the training, but still keep all affixes,
you will need to add all the affixes explicitly to this list. You must use the +c
switch when using +e.
No specific file of affixes and stems isused for training. (This is the default, unless
a tags.cut file is present.)
+mN load the disambiguation matrices into memory (about 700K)
N=0 no matrix training
N=2 training with matrix of size 2
N=3 training with matrix of size 3
N=4 training with matrix of size 4 (default)
append errors output to file F (default: send it to screen)
This switch has three forms
N=0 default log listing mismatches between the %trn and %mor line.
N=1 similar output in a format designed more for developers.
N=2 complete output of all date, including both matches and mismatches
include tier code S
exclude tier code S
+/-t#Target_Child - select target child's tiers
+/[email protected]="*|Mother|*" - select mother's tiers
Part 3: Morphosyntax
use syntactic category suffixes to deal with stem compounds
When using the default switch form of the error log, lines that begin with @ indicate that
the %trn and %mor had different numbers of elements. Lines that do not begin with @
represent simple disagreement between the %trn and the %mor line in some category
assignment. For example, if %mor has pro:dem^pro:exist and %trn has co three times.
Then +s0 would yield: 3 there co (3 {1} pro:dem (2} pro:exist).
By default, POSTTRAIN uses all the affixes in the language and none of the stems. If you
wish to change this behavior, you need to create a file with your grammatical names for
prefixes and suffixes or stem tags. This file can be used by both POSTTRAIN and POST.
However, you may wish to create one file for use by POSTTRAIN and another for use by
The English POST disambiguator currently achieves over 95% correct disambiguation. We
have not yet computed the levels of accuracy for the other disambiguators. However, the
levels may be a bit better for inflectional languages like Spanish or Italian. To train the
POST disambiguator, we first had to create a hand-annotated training set for each language.
We created this corpus through a process of bootstrapping. Here is the sequence of basic
steps in training.
1. First run MOR on a small corpus and used the Esc-2 hand disambiguation process
to disambiguate.
2. Then rename the %mor line in the corpus to %trn.
3. Run MOR again to create a separate %mor line.
4. Run POSTTRAIN with this command: posttrain +c +o0err.cut +x *.cha
5. This will create a new post.db database.
6. You then need to go through the 0errors.cut file line by line to eliminate each
mismatch between your %trn line and the codes of the %mor line. Mismatches
arise primarily from changes made to the MOR codes in between runs of MOR.
7. Before running POST, make sure that post.db is in the right place. The default
location is in the MOR library, next to ar.cut and cr.cut. However, if post.db is not
there, POST will look in the working directory. So, it is best to make sure it is in
the MOR library to avoid confusion.
8. Disambiguate the MOR line with: post *.cha +1
9. Compare the results of POST with your hand disambiguation using: trnfix *.cha
When using TRNFIX, sometimes the %trn will be at fault and sometimes %mor will be at
fault. You can only fix the %trn line. To fix the %mor results, you just have to keep on
compiling more training data by iterating the above process. As a rule of thumb, you
eventually want to have at least 5000 utterances in your training corpus. However, a corpus
with 1000 utterances will be useful initially.
During work in constructing the training corpus for POSTTRAIN, you will eventually
bump into some areas of English grammar where the distinction between parts of speech
is difficult to make without careful specification of detailed criteria. We can identify three
areas that are particularly problematic in terms of their subsequent effects on GR
Part 3: Morphosyntax
(grammatical relation) identification:
1. Adverb vs. preposition vs. particle. The words about, across, “after”, away, back,
down, in, off, on, out, over, and up belong to three categories: ADVerb, PREPosition
and ParTicLe. In practice, it is usually impossible to distinguish a particle from an
adverb. Therefore, we only distinguish adverbs from prepositions. To distinguish
these two, we apply the following criteria. First, a preposition must have a
prepositional object. Second, a preposition forms a constituent with its noun phrase
object, and hence is more closely bound to its object than an adverb or a particle.
Third, prepositional phrases can be fronted, whereas the noun phrases that happen to
follow adverbs or particles cannot. Fourth, a manner adverb can be placed between
the verb and a preposition, but not between a verb and a particle.
2. Verb vs. auxiliary. Distinguishing between Verb and AUXiliary is especially tricky
for the verbs be, do and have. The following tests can be applied. First, if the target
word is accompanied by a nonfinite verb in the same clause, it is an auxiliary, as in I
have had enough or I do not like eggs. Another test that works for these examples is
fronting. In interrogative sentences, the auxiliary is moved to the beginning of the
clause, as in have I had enough? and do I like eggs? whereas main verbs do not move.
In verb-participle constructions headed by the verb be, if the participle is in the
progressive tense (John is smiling), then the head verb is labeled as an AUXiliary,
otherwise it is a Verb (John is happy).
3. Copula vs. auxiliary. A related problem is the distinction between v:cop and aux for
the verb to be. This problem arises mostly when the verb is followed by the past
participle, as in I was finished. For these constructions, we take the approach that the
verb is always the copula, unless there is a by phrase marking the passive.
4. Communicators. COmmunicators can be hard to distinguish imperatives or locative
adverbs, especially at the beginning of a sentence. Consider a sentence such as there
you are where there could be interpreted as either specifying a location vs. there is a
car in which there is pro:exist.
This tool enables you to modify the Brill rules of a database. There are these options:
+dF use POST database file F (default is eng.db).
specify name of file (F) containing actions that modify rules.
force creation of Brill's rules.
+lm reduce memory use (but increase processing time).
Part 3: Morphosyntax
10 GRASP – Syntactic Dependency Analysis
This chapter, with contributions from Eric Davis, Shuly Wintner, Brian MacWhinney,
Alon Lavie, Andrew Yankes, and Kenji Sagae, describes a system for coding syntactic
dependencies in the English TalkBank corpora. This chapter explains the annotation
system and describes each of the grammatical relations (GRs) available for tagging
dependency relations.
10.1 Grammatical Relations
GRASP describes the structure of sentences in terms of pairwise grammatical relations
between words. These grammatical relations involve two dimensions: attachment and
valency. In terms of attachment, each pair has a head and a dependent. These dependency
relations are unidirectional and cannot be used to represent bidirectional relations. Along
the valency dimension, each pair has a predicate and an argument. Each dependency
relation is labeled with an arc and the arc has an arrow which points from the predicate to
argument. Valency relations open slots for arguments. In English, modifiers (adjectives,
determiners, quantifiers) are predicates whose arguments are the following nouns. In this
type of dependency organization, the argument becomes the head. However, in other
grammatical relations, the predicate or governor is the head and the resultant phrase takes
on its functions from the predicate. Examples of predicate-head GRs include the
attachment of thematic roles to verbs and the attachment of adjuncts to their heads.
Here is an example of the coding of the sentence the big dog chased five cats for
*TXT: the big dog chased five cats.
%mor: det|the adj|big n|dog v|chase-PAST quant|five n|cat-PL.
1|3|DET 2|3|MOD 3|4|SUBJ 4|0|ROOT 5|6|QUANT 6|4|OBJ
This notation can be described in this way:
1. The determiner the is the first item and it attaches to the third item dog. Here the
determiner is the predicate and the dependent. The GR here is DET or determination.
2. The adjective big is a predicate that attaches as a dependent of dog. The GR here is
MOD or modification.
3. The noun dog is the head of the phrase the big dog and it attaches as a dependent
subject or SUBJ of the predicate chased. Here we ignore the attachment of the suffix
–ed to the verb.
4. The verb chased is the root of the clause. It attaches to the zero position which is the
“root” of the sentence.
5. The quantifier five attaches to the noun cats through the QUANT relation.
6. The noun cats attaches as a dependent to the verb chased through the OBJ or object
The following diagram describes the sentence We eat the cheese sandwich graphically
through arcs with arrowheads instead of through the numbering system:
Part 3: Morphosyntax
This picture is equivalent to this notation in the numbering system:
*TXT: we eat the cheese sandwich.
%mor: pro|we v|eat det|the n|cheese n|sandwich .
%gra: 1|2|SUBJ 2|0|ROOT 3|5|DET 4|5|MOD 5|2|OBJ
The creation of these %gra notations, either by hand or by machine, depends on three levels
of representation.
1. Words must be consistently coded for the correct part of part of speech on the %mor
line. For English, the relevant categories are given in the MOR grammars. The major
categories are adjective, adverb, pronoun, noun, and verb. However, there are many
additional subcategories. Compounds are treated as single units given in their main
part of speech by MOR.
2. GRASP makes use of a large set of grammatical relations (GRs), given below.
3. GRs apply to phrases or clusters of words connected through dependency relations.
The most import phrasal types are: noun phrase, verb phrase, absolute phrase, gerund
phrase, auxiliary phrase, and infinitival phrase. To achieve attachment to heads,
phrases must be represented by one of their component lexical items. In some
structures this can be a relativizer or conjunction; in others, it can be the main verb of
the subordinate clause.
The following is a comprehensive list of the grammatical relations in the GRASP
annotation scheme. Example GRs as well as relations to similar GRs are provided. In this
annotation scheme, C refers to clausal and X refers to non-finite clausal. This list is divided
into relations with the predicate as head and relations with the argument as head. In the
examples, the dependent is marked in italics.
10.2 Predicate-head relations
First, we list the relations in which the dependent attaches to a head that serves as the
predicate. In many of these relations, the head is the verb. The combination of a verb with
its arguments, including the SUBJ argument, constitutes a verb phrase.
1. SUBJect identifies the subject of clause, when the subject itself is not a clause.
Typically, the head is the main verb and the dependent is a nominal. Ex: You eat
with your spoon.
2. ClausalSUBJect = CSUBJ identifies the finite clausal subject of another clause. The
head is the main verb, and the dependent is the main verb of the clausal subject. An
Part 3: Morphosyntax
alternative analysis of these structures would treat the subordinator “that” as the
head of the CSUBJ. Ex: That Eric cried moved Bush.
3. OBJect identifies the first object or direct object of a verb. The head is the main
verb, and the dependent is a nominal or a noun that is the head of a nominal phrase.
A clausal complement relation should be denoted by COMP or XCOMP
(depending on whether the clausal complement is finite or non-finite, see below),
not OBJ or OBJ2. Ex: You read the book.
4. OBJect2 = OBJ2 identifies the second object or indirect object of a ditransitive
verb, when not introduced by a preposition. The head is a ditransitive verb, and the
dependent is a noun (or other nominal). The dependent must be the head of a
required non-clausal and nonprepositional complement of a verb (head of OBJ2)
that is also the head of an OBJ relation. Ex: He gave you your telephone. When the
indirect object is in a prepositional phrase, it is just coded as a prepositional phrase.
5. COMPlement identifies a finite clausal complement of a verb. The head is the main
verb of the matrix clause, and the dependent is the main verb of the clausal
complement. Ex: I think that was Fraser.
6. XCOMPlement identifies a non-finite clausal complement of a verb. The head is
the main verb of the matrix clause, and the dependent is the main verb of the clausal
complement. The XCOMP relation is only used for non-finite clausal complements,
not predicate nominals or predicate adjectives (see PRED). Ex: You’re going to
stand on my toe. I told you to go. Eve, you stop throwing the blocks.
7. PREDicate identifies a predicate nominal or predicate adjective of verbs such as be
and become. The head is the verb. PRED should not be confused with XCOMP,
which identifies a non-finite complement of a verb (some syntactic formalisms
group PRED and XCOMP in a single category). Ex: I’m not sure. He is a doctor.
8. ClausalPREDicate = CPRED identifies a full clause (finite or non-finite) that serves
as the predicate nominal of verbs such as be and become. The head is the verb of
the main or matrix clause. The dependent is the verb of the predicate clause. If there
is a relativizer, it attaches to the verb of th embedded clause through the LINK
relation. Ex: This is how I drink my coffee. My goal is to win the competition.
9. ClausalPrepositionalOBJect = CPOBJ indentifies a full clause that serves as the
object of a preposition. The verb of the clause attaches to the preposition which is
the head. Here, again, the relativizer attaches to the verb of the subordinate clause
through the LINK relation. Ex: I’m confused about what you are asking.
10. ClausalOBJect = COBJ indentifies a full clause that serves as OBJ. The verb of
the object clause attaches to the main verb which is the head. Here, again, the
relativizer attaches to the verb of the subordinate clause through the LINK relation.
Ex: I remember what you said.
11. PrepositionalOBJect = POBJ is the relation between a preposition and its object.
The head is a preposition, and the dependent is typically a noun. The traditional
treatment of the prepositional phrase views the object of the preposition as the head
of the prepositional phrase. However, we are here treating the preposition as the
head, since the prepositional phrase then participates in a further JCT relation to a
head verb or a NJCT relation to a head noun. Ex: You want to sit on the stool?
12. SeRiaL identifies serial verbs such as go play and come see. In English, such verb
sequences start with either come or go. The initial verb is the dependent, and the
Part 3: Morphosyntax
following verb is the head and typically the root of the sentence. Ex: Come see if
we can find it. Go play with your toys over there. This relation can also be used
when children overuse the pattern in forms with omitted infinitives such as want
see, try go. It is not used with auxiliary have, although sometimes it seems that it
could be.
10.3 Argument-head relations
Relations in which the arguments (rather than the predicates) serve as the heads include
relations of adjunction and modification.
1. adJunCT = JCT identifies an adjunct that modifies a verb, adjective, or adverb. This
grammatical relation covers a wide variety of structures whose exact form is best
understood by noting the specific parts of speech that are involved. In these, the
adjunct is the predicate, since it opens a valency slot for something to attach to.
The head of JCT is the verb, adjective or adverb to which the JCT attaches as a
dependent. The dependent is typically an adverb or a preposition (in the case of
phrasal adjuncts headed by a preposition, such as a prepositional phrase).
Occasionally, a locative noun may function as an adjunct, as with way in he was
going all the way home. Adjuncts are optional, and carry meaning on their own
(and do not change the basic meaning of their JCT heads). Verbs requiring a
complement describing location may be treated as prepositional objects, in which
case the IOBJ relation applies (see above). Ex: That’s much better. He ran with a
limp. That’s really big.
2. ClausalconJunCT = CJCT identifies a finite clause that adjoins coordinatively to a
verb, adjective, or adverb head. The dependent is the main verb of the subordinate
clause and this clause attaches to the root verb of the main clause. The conjunction
uses the LINK relation and attaches to the verb of the subordinate clause. Ex: We
can’t find it, because it is gone. When two clauses are combined with and CJCT is
the link between the main verb of the second clause and the first. However, this
relation should not be used if the subject is not repeated, as in John runs 5 miles,
and swims one every day. Instead, this is a case of a coordinated verb.
3. XadJunCT = XJCT identifies a non-finite clause that attaches to a verb, adjective,
or adverb. The dependent is typically the main verb of a non-finite subordinate
clause. There is usually no conjunction for these. Ex: Running to the carriage, she
lost her slipper. Note: this construction can be confused with ones in which a PP
or adverb intervenes between an auxiliary and its participle, as in she’s outside
sleeping in the carriage. However, when the main verb is a copula, as in there’s a
man sleeping in the car, then XJCT is appropriate.
4. Nominal adJunCT = NJCT identifies the head of a complex NP with a prepositional
phrase attached as an adjunct of a noun. Note that, if a prepositional phrase attaches
to a verb, the relation is JCT and not NJCT. In a sense, this relation is a cross
between JCT and MOD. Ex: The man with an umbrella arrived late.
5. MODifier identifies a non-clausal nominal modifier. This should not be confused
with the part of speech code with the same name for modals. The link of modals to
the main verb is coded as AUX. The head is a noun, and the dependent is typically
an adjective or another noun, including a possessive. Ex: Would you like grape
juice? That’s a nice box. That’s John’s stick.
Part 3: Morphosyntax
6. POSTMODifier identifies a postposed nominal modifier. The head is a noun, and
the dependent is a following adjective. Ex: I will dig a hole bigger than a foot.
Often these express resultative relations, as in I painted the barn red. Sometimes
they occur after intervening material, as in He pulled her out of the water safe.
7. POSSessive is used for the relation between the English possessive suffix and the
head noun. The head is the noun. The introduction of this relation is required to
avoid problems in processing the verb clitics that take the same phonological shape.
Ex: John’s book.
8. APPositive identifies the relation between an appositive phrase and its head. Ex:
Barack Obama, President of the United States. It can also be used to link a
topicalized noun to a following pronoun. Ex: John, he really likes waffles.
9. ClausalMODifier = CMOD identifies a finite clause that is a nominal modifier (such
as a relative clause) or complement. The head is a noun, and the dependent is
typically a finite verb. Ex: Here are the grapes I found. This relation can also be
used when the head is an adjective. Ex: He was happy he found the girl.
10. XMODifier identifies a non-finite clause that is a nominal modifier (such as a
relative clause) or complement. The head is a noun, and the dependent is typically
a non-finite verb. Ex: It’s time to take a nap. I saw a man running away.
11. DETerminer identifies the relation between a determiner and its head noun.
Determiners include the, a, as well as (adjectival) possessives pronouns (my, your,
etc) and demonstratives (this, those, etc), but not quantifiers (all, some, any, etc; see
QUANT below). Typically, the head is a noun and the dependent/governor is a
determiner. In cases where a word that is usually a determiner does not have a head,
there is no DET relation. Ex: I want that cookie.
12. QUANTifier identifies a nominal quantifier, such as three, many, and some.
Typically, the head is a noun, and the dependent is a quantifier, or sometimes an
adverb. In cases where a quantifier has no head, there is no QUANT relation. In
English, the MOD, DET, and QUANT relations have largely the same syntax.
However, within the noun phrase, we occasionally see that they are ordered as
DET+QUANT+MOD+N. Ex: These are my three ripe bananas.
13. PostQuantifier = PQ is the relation between a postquantifier and the preceding head
nominal. Ex: We both arrived late.
14. AUXiliary identifies an auxiliary or modal of a main verb. The head is a verb, and
the dependent is an auxiliary (such as be or have) or a modal (such as can or should).
Ex: Can you do it?
15. NEGation identifies verbal negation. When the word not (contracted or not) follows
an auxiliary or modal (or sometimes a verb), it is the dependent in a NEG relation
(not JCT), where the auxiliary, modal or verb (in the absence of an auxiliary or
modal) is the head. Ex: Mommy will not read it.
16. INFinitive identifies the relation between the infinitival particle (to) and the verb to
which it attaches. The head is a verb, and the dependent is always to. Ex: He’s going
to drink the coffee.
17. LINK identifies the relation between a complementizer (that), relativizer (who,
which) or subordinate conjunction (including and) and the verb in the subordinate
clause to which it attaches. The verb of the subordinate clause attaches to the main
verb in a CJCT, CMOD, COBJ, CPOBJ, CPRED, or COMP relation. Ex: Wait until
Part 3: Morphosyntax
the noodles are cool.
18. TAG is the relation between the finite verb of a tag question and the root verb of
the main clause. Ex: You know how to count, don’t you? English and Portuguese
have this structure, but many other languages do not.
10.4 Extra-clausal elements
In these relations, the dependent is a clausal modifier. These should be linked to the root.
1. COMmunicator identifies a communicator (such as hey, okay, etc) or a vocative
when it occurs inside an utterance. When these occur at the beginning or end of
utterances, then use BEG and END instead. The head of COM is the ROOT.
2. BEGin identifies an initial clause-external element, such as a vocative or topic. The
head of BEG is “0”. Ex: Eve, are you coming? BEG is marked in the main line
with a following ‡ mark that is coded on the %mor line as beg|begp. The BEGP
relation is linked to BEG.
3. END identifies a final clause-external postposed element, including sentence final
particles, final vocatives, final interactionals, and single word tags like right? Like
COM, but unlike BEG, the head of the END is the ROOT. This is done, because
the parser works better in this way. END is marked in the main line with a preceding
„ mark that is coded on the %mor line as end|endp. Ex: Some more cookies, Eve?
4. INCompleteROOT identifies a word that serves as the root of an utterance, because
the usual root forms (verbs in English) are missing. This form could be a single
word by itself (adverb, communicator, noun, adjective) or a word with additional
modifiers, such as the noun dog in the big dog, when it occurs by itself without a
verb. It may appear that there could be more than one of these in an utterance, as
in well, sure. However, in this case, well should be marked as a CO that is
dependent on sure.
5. OMission is used when ellipsis or omission leaves determiners or other modifiers
without heads. They should then be attached to the most local predicate using the
OM relation.
10.5 Cosmetic relations
There are several relations that are just used during transcription to assist in the accurate
training of the GRASP tagger:
1. PUNCTuation is the relation between the final punctuation mark and ROOT.
2. LP (local punctuation) is the relation between quotes or commas and overall
structure. Commas delimiting clauses should be linked to the root. Each comma in
an ENUM series should be linked to the previous word. Beginning and ending
quote marks should be linked to “0”.
3. BEGP is the relation between the ‡ mark and the BEG.
4. ENDP is the relation between the „ mark and the END.
5. ROOT This is the relation between the topmost word in a sentence (the root of the
dependency tree) and the LeftWall or “0”. The topmost word in a sentence is the
word that is the head of one or more relations, but is not the dependent in any
relation with other words (except for the LeftWall).
Part 3: Morphosyntax
Series relations. Some additional relations involve processes of listing, coordination, and
classification. In some of these, the final element is the head and the initial elements all
depend on the final head. However, in coordinations, we take the first element as the head.
1. NAME identifies a string of proper names such as Eric Davis and New York
Central Library. The initial name is the dependent, and the following name is the
head. Ex: My name is Tom Jones.
2. DATE identifies a date with month and year, month and day, or month, day, and
year. Examples include: October 7, 1980 and July 26. For consistency with
compounds and NAME, we regard the final element in these various forms as the
head. Ex: October seventh nineteen ninety.
3. ENUMeration involves a relation between elements in a series without any
coordination based on a conjunction (and, but, or). The series can contain letters,
numbers, and nominals. The head is the first item in the series, and all the other
items in the enumeration depend on this first word. Ex: one, two, three, four.
4. CONJ involves a relation between a coordinating conjunction and the clustr of
preceding conjoined items that has been combined with ENUM. For example, in
the phrase I walk, jump, and run, the items walk and jump are combined by ENUM
so that walk is the head. The conjunction and then attaches to walk with the CONJ
relation. The resultant phrase walk, jump and is then further linked by the COORD
relation to the final element ran. Ex: I walk, jump, and run.
5. COORD involves an attachment of a final coordinated element to the conjunction.
For example, in the sentence I walk, jump, and run, the verb run is attached to and.
Note: When the word and functions as a subordinating conjunction, it is treated as
having a LINK relation to the verb of its clause. In this case, it is not conjoining
phrases but whole clauses. In the neither X nor Y and either X or Y constructions,
the words neither, nor, either, and or are treated as coord by MOR and each coord
words links to the following content word (noun, verb, or adjective) with COORD.
Then the first content word links to the second with ENUM.
MEGRASP is the CLAN command that creates a %gra line based on the forms in the %mor
line. MEGRASP uses a maximum entropy classifier system encoded in the megrasp.mod
database file to create a dependency parsing based on binary grammatical relations. To
visualize these dependencies graphically, you can triple click on a %gra line and CLAN
will run a web service that produces labelled graph structure on your computer screen. The
figures in the next chapter are screengrabs taken from those displays.
To create a new megrasp.mod database file, you can run MEGRASP in training mode used
the –t switch. In that case, your training corpus must contain a gold standard %grt tier to
control training. The program uses these options:
-e :
evaluate accuracy (input file must contain gold standard GRs)
-t :
training mode (parser runs in parse mode by default)
-iN: number of iterations for training ME model (default: 300)
-cN: inequality parameter for training ME model (default: 0.5).
+fS: send output to file (program will derive filename)
+re: run program recursively on all sub-directories.
Part 3: Morphosyntax
11 Building a training corpus
Once a megrasp.mod file has been created by running MEGRASP in training mode, the
creation of the %gra line follows automatically in the chain of MOR-POSTPOSTMORTEM-MEGRASP. However, to produce accurate tagging it is important to
have properly tagged GRs in the training corpus. To do this accurately involves making
consistent decisions for certain GRs that are easy to confuse. This chapter provides
illustrations that can help in making these choices between competing GR assignments.
11.1 OBJ and OBJ2
OBJect identifies the direct object or first object of a verb. (As with SUBJ, this should be
a noun phrase. If it’s a clause, see section 13.)
The head of OBJ is the verb for which it is acting as an object.
OBJect2 identifies the indirect object or second object of a verb, if it is not introduced by
a preposition.
The head of OBJ2, like OBJ, is the verb for which it’s acting as an indirect object.
*PAR: the worker pushes the box .
%mor: art|the n|work&dv-AGT v|push-3S art|the n|box .
%gra: 1|2|DET 2|3|SUBJ 3|0|ROOT 4|5|DET 5|3|OBJ 6|3|PUNCT
This sentence is structured almost exactly like example 1a except for the presence of an
object. box ties to pushes as its OBJ.
*PAR: the child gave the dog a treat .
%mor: art|the n|child v|give-PAST art|the n|dog art|a n|treat .
%gra: 1|2|DET 2|3|SUBJ 3|0|ROOT 4|5|DET 5|3|OBJ2 6|7|DET 7|3|OBJ 8|3|PUNCT
The verb gave takes three arguments here. Other than its subject child, these are dog and
treat. treat is the direct object of gave, whereas dog is the indirect object (the beneficiary
of the action), and is not introduced by a preposition. So, treat and dog are identified as
OBJ and OBJ2, respectively.
Part 3: Morphosyntax
11.2 3. JCT and POBJ
adJunCT = JCT identifies an adjunct that modifies a verb, adjective, or adverb (but not a
noun – see section 4). This covers a very wide range of grammatical structures, but the
word identified as a JCT is typically a preposition or adverb. JCTs very commonly head
up prepositional phrases. (A JCT may also be a noun in some cases, such as way in she
walked all the way home.)
The head of JCT is the verb, adjective, or adverb it modifies.
Prepositional OBJect = POBJ identifies the (non-clause) object of a preposition, such as
town in the phrase around town. (For a clausal POBJ, see section 13.)
The head of POBJ is the preposition on which the object depends.
Note: JCT and POBJ also work together to describe phrasal verbs, despite the fact that their
function is not truly adjunctive. For example, in the sentence he took out the trash, out is
tied to took as a JCT and trash is tied to out as a POBJ.
This isn’t a strictly accurate parse of the sentence, since out the trash is not an optional
constituent here; rather, the trash would be more properly understood as an OBJ to the
phrasal verb took out. However, under current GRASP ruling, it’s more convenient to treat
this as a JCT-POBJ construction.
Note: When a prepositional phrase acts as the predicate in a be construction, treat it as a
JCT (and a POBJ, if there is one). A simple example would be she is in the boathouse.
You might find it counterintuitive to tie in to is as a JCT there, since in the boathouse is,
again, not an optional constituent. (As you’ll see in the next section, it could be tempting
to call in the boathouse a PRED.) It’s more consistent to treat prepositional phrases this
way, though, even when they’re not adjunctive.
*PAR: the child gave a treat to the dog .
%mor: art|the n|child v|give-PAST art|a n|treat prep|to det|the n|dog .
%gra: 1|2|DET 2|3|SUBJ 3|0|ROOT 4|5|DET 5|3|OBJ 6|3|JCT 7|8|DET 8|6|POBJ 9|3|PUNCT
This is nearly the same sentence as example 2b, except that dog is now introduced by a
preposition (and, accordingly, the arguments are rearranged). Here, to the dog is treated as
an adjunct to the verb. Thus to ties as a JCT to gave, and dog ties as a POBJ to to.
Part 3: Morphosyntax
*PAR: I really messed up that test.
%mor: pro:sub|I adv|real&dadj-LY v|mess-PAST prep|up det|that n|test .
%gra: 1|3|SUBJ 2|3|JCT 3|0|ROOT 4|3|JCT 5|6|DET 6|4|POBJ 7|3|PUNCT
First, note the adverb really tying as a JCT to the verb it modifies, messed. Otherwise, the
purpose of this example is demonstrating how to diagram a phrasal verb in GRASP. As
you can see, the way to handle the phrasal verb messed up is to treat up that test as an
adjunct to messed, even though that’s a slight compromise in accuracy.
*PAR: Peter is outside the tower .
%mor: n:prop|Peter cop|be&3S prep|outside art|the n|tower .
%gra: 1|2|SUBJ 2|0|ROOT 3|2|JCT 4|5|DET 5|3|POBJ 6|2|PUNCT
A straightforward subject-be-PP construction. outside the tower is treated as a JCT to is in
order to maximize consistency in the treatment of prepositional phrases, allowing for the
fact that this leaves the copula is in the strange position of not appearing to take any
11.3 PRED and NJCT
PREDicate identifies a nominal or adjectival predicate. This is usually an argument of the
verb be, although other verbs such as become and grow (in the sense of I grow weary of
your complaints) can also take a predicate.
The head of PRED is the verb of which it is an argument.
Nominal adJunCT = NJCT identifies an adjunct that modifies a noun rather than a verb,
adjective, or adverb. Otherwise it behaves the same as JCT, able to take a POBJ and so on.
The head of NJCT is the noun it modifies.
PRED also participates along with SUBJ in there’s-an-X constructions. Treat there as an
existential pronoun in these constructions, tying it to the be verb as a SUBJ. Then tie the
introduced noun phrase to the be verb as a PRED. (Do the same for here’s-an-X.)
Linguistically, other solutions are conceivable, but this is the most reliable for GRASP
Part 3: Morphosyntax
*PAR: you became wiser .
%mor: pro:sub|you v|become-PAST adj|wise-CP .
%gra: 1|2|SUBJ 2|0|ROOT 3|2|PRED 4|2|PUNCT
Here is a simple adjectival predicate.
*PAR: she is a friend of my sister .
%mor: pro:sub|she cop|be&3S art|a n|friend prep|of pro:poss:det|my n|sister .
%gra: 1|2|SUBJ 2|0|ROOT 3|4|DET 4|2|PRED 5|4|NJCT 6|7|DET 7|5|POBJ 8|2|PUNCT
A nominal predicate, with an NJCT attached. of my sister is an adjunct phrase here, but
since it modifies a noun, we class it as an NJCT rather than a regular JCT.
*PAR: there's a ghost in the cloakroom .
%mor: pro:exist|there~cop|be&3S art|a n|ghost prep|in art|the n|+n|cloak+n|room .
%gra: 1|2|SUBJ 2|0|ROOT 3|4|DET 4|2|PRED 5|4|NJCT 6|7|DET 7|5|POBJ 8|2|PUNCT
An example of there’s-an-X. We treat there as the SUBJ and ghost as a PRED. Notice that
there’s some mild structural ambiguity as to where the prepositional phrase should tie –
you might read this sentence as either “In the cloakroom, a [ghost] exists” or “A [ghost in
the cloakroom] exists”. What is shown above is the latter interpretation, where in ties to
ghost as an NJCT; the other interpretation would suggest tying in to be as a regular JCT.
When such ambiguity arises, it’s a safe rule of thumb to pick the solution that involves
tying closer words together, which is an easier process to generalize.
Part 3: Morphosyntax
11.4 AUX and NEG
AUXiliary identifies an auxiliary or modal of a main verb, such as can or should. The head
of AUX is the verb which it serves. Even though an AUX typically takes over tense
information from the verb it serves, that main verb is still considered the ROOT.
Note that when a SUBJ and an AUX are present along with a main verb, both the SUBJ
and the AUX individually tie to the main verb. (The SUBJ doesn’t tie to the AUX.)
Relatedly, when multiple AUXes are present (could have guessed), each again ties
individually to the main verb, for convenience. (None of the AUXes tie to each other.)
NEGation identifies a verbal negator; i.e. not (including its contracted form -n’t). In
modern English, the presence of a NEG is nearly always accompanied by an AUX, a form
of be as a main verb (she isn’t friendly), or sometimes a participle (consider not telling
anyone that). NEG can also in some cases negate a relativizer (see section 10 for details on
The head of NEG is the AUX, verb, or relativizer it negates.
Example 5a
*PAR: the lawyer can help us .
%mor: art|the n|lawyer mod|can v|help pro:obj|us .
%gra: 1|2|DET 2|4|SUBJ 3|4|AUX 4|0|ROOT 5|4|OBJ 6|4|PUNCT
A simple auxiliary construction. While can here carries the tense information, the ROOT
of the sentence is still the main verb, help. As mentioned above, lawyer ties directly to help
as the SUBJ (it does not tie to can, the AUX).
Example 5b
*PAR: don't you enjoy the theatre ?
%mor: mod|do~neg|not pro:sub|you v|enjoy art|the n|theatre ?
%gra: 1|4|AUX 2|1|NEG 3|4|SUBJ 4|0|ROOT 5|6|DET 6|4|OBJ 7|4|PUNCT
This demonstrates an AUX-leading question as well as a negated AUX. Structurally it’s no
different from you don’t enjoy the theatre, just rearranged. The NEG ties to the AUX, not
the main verb enjoy.
Part 3: Morphosyntax
11.5 MOD and POSS
MODifier identifies a non-clausal nominal modifier which precedes the noun it modifies.
This might generally be an adjective, a possessive, or another noun (as in the cookie factory,
where cookie is a MOD to factory). The head of MOD is the noun it modifies.
When multiple MODs modify the same noun, each ties to that noun separately; they don’t
stack on each other. (But MODs can modify other MODs. In the oatmeal cookie factory,
oatmeal modifies cookie, and cookie modifies factory. This is structurally unlike big bronze
sculpture, where both big and bronze individually modify sculpture.) Likewise, if the
modified noun is targeted by a determiner, that DET ties to the noun itself, not any MOD.
MODs sharing a head do not interact either with each other or with DETs.
POSSessive identifies the relation between the English possessive suffix and the noun it
affixes to, such as -’s in John’s book.
The head of POSS is the noun it affixes to (John in the given example, not book). The
affixed noun then ties as a MOD to the noun it specifies (so John ties to book).
*PAR: then we'll find a new apartment .
%mor: adv:tem|then pro:sub|we~mod|will v|find art|a adj|new n|apartment .
%gra: 1|4|JCT 2|4|SUBJ 3|4|AUX 4|0|ROOT 5|7|DET 6|7|MOD 7|4|OBJ 8|4|PUNCT
A simple MOD construction. As described above, both the DET a and the MOD new tie
individually to apartment (a does not tie to new).
*PAR: this is Sarah's greatest piece in the exhibition .
%mor: pro:dem|this cop|be&3S n:prop|Sarah~poss|s adj|great&SP n|piece
prep|in art|the n|exhibition .
%gra: 1|2|SUBJ 2|0|ROOT 3|6|MOD 4|3|POSS 5|6|MOD 6|2|PRED 7|6|NJCT 8|9|DET
9|7|POBJ 10|2|ROOT
POSS in action. The -’s of Sarah’s is treated as its own unit and ties to Sarah as a POSS,
while Sarah ties to piece as a MOD (greatest also ties to piece as a separate MOD).
11.6 CONJ, and COORD
Part 3: Morphosyntax
CONJunction identifies a coordinating conjunction such as and, but, and or, when it is
serving to conjoin two or more equivalent phrases (fire and brimstone). The head of CONJ
is the first of the conjoined phrases. (Even when more than two items are being conjoined,
every CONJ will tie to the first item. They don’t form a chain, with each CONJ tying to
the directly preceding item.)
Some conjunctions occur at the beginning of a sentence and apply over the entire
sentence, as in but the film adaptation was great. For these, use LINK rather than CONJ
(see section 10).
COORDination identifies any element but the first in a series of conjoined equivalent
phrases. The head of COORD is the CONJ conjoining it.
For example, in the phrase bold and eloquent woman, only the first adjective, bold, ties
to woman as a MOD. and ties to bold as a CONJ, while eloquent ties to and as a COORD.
The overall effect is to unite bold and eloquent into one constituent headed by its first
element, bold.
CONJ and COORD are not used when conjoining entire clauses that have non-matching
subjects, like I popped the cork and champagne sprayed out. Instead see section 14.
*PAR: she scrimped and saved and bought a car .
%mor: pro:sub|she v|scrimp-PAST conj|and v|save-PAST conj|and v|buy-PAST art|a n|car .
%gra: 1|2|SUBJ 2|0|ROOT 3|2|CONJ 4|3|COORD 5|2|CONJ 6|5|COORD 7|8|DET 8|6|OBJ
A simple CONJ-COORD construction. There are three conjoined verb phrases. The first
one, scrimped, will serve as the ROOT, while the other two, saved and bought, will be
treated as COORDs. Each of these two verbs ties to the conjunction and preceding it; then,
each and ties to the first verb, scrimped. As stated above, the three verbs don’t form a chain.
Therefore, the second and doesn’t tie to saved.
11.7 ENUM and LP
ENUMeration identifies a non-initial serial element which is not explicitly conjoined by
its own CONJ. It’s very common in English to list three or more items while only explicitly
naming one CONJ, as in the lion, the tiger, and the panther; or even no CONJ at all, as in
we laughed, schmoozed, had a great time.
These are cases where the ENUM class is called for. In the first example, and the panther
ties to lion just like a regular CONJ-COORD pair (and ties to lion as a CONJ; panther ties
to and as a COORD). However, this leaves the tiger, which has no CONJ. Instead, tiger
ties directly to lion as an ENUM. In the second example, where no CONJs are present at
all, both of the verbs schmoozed and had tie to laughed as ENUMs.
Part 3: Morphosyntax
Again, as in the previous section, noticee that in both examples, everything is ultimately
tying to the first item in the series (rather than creating a chain where each item ties to the
one directly before it).
Local Punctuation = LP identifies commas and quotes. When a comma delimits items in
a series, each comma ties to the item preceding it. When a comma delimits entire clauses,
it ties to the root. Quote marks either at the beginning or the end of utterances tie to 0, the
Note: COORD and ENUM also participate in the formulations either/neither X or/nor Y
and as X as Y. In these cases, the conjoining terms either/neither/or/nor/as are tagged
COORD, and each one ties to the element it precedes. Then the second element ties to the
first element as an ENUM. The overall effect is to unite the entire formulation into one
constituent headed by its first element.
*PAR: Ron is helpful , considerate , and professional .
%mor: n:prop|Ron cop|be&3S adj|helpful cm|cm adj|considerate cm|cm conj|and adj|professional
%gra: 1|2|SUBJ 2|0|ROOT 3|2|PRED 4|3|LP 5|3|ENUM 6|5|LP 7|3|CONJ 8|7|COORD
Here considerate is not connected to helpful by an explicit CONJ, so, instead, it ties to
helpful by the ENUM relation. and professional, on the other hand, exhibits a typical
CONJ-COORD relation to helpful. Of the two commas in the sentence, each one ties to the
list item preceding it.
*PAR: be as quiet as a mouse .
%mor: v|be prep|as adj|quiet prep|as art|a n|mouse .
%gra: 1|0|ROOT 2|3|COORD 3|1|PRED 4|6|COORD 5|6|DET 6|3|ENUM 7|1|PUNCT
An example of as X as Y. The entire phrase as quiet as a mouse is bound into one
constituent, with the PRED quiet as its head. This is accomplished by tying each as to its
corresponding item as a COORD, and then tying mouse to quiet as an ENUM.
Part 3: Morphosyntax
*PAR: neither the pink one nor the blue one fits well .
%mor: conj|neither art|the adj|pink n|one conj|nor art|the adj|blue n|one v|fit-3S adv|well .
%gra: 1|4|COORD 2|4|DET 3|4|MOD 4|9|SUBJ 5|8|COORD 6|8|DET 7|8|DET 8|4|ENUM
9|0|ROOT 10|9|JCT 11|9|PUNCT
An example of neither X nor Y. As in example 8c above, the goal is to bind together neither
the pink one nor the blue one into one constituent headed by [pink] one, which can then
act as the SUBJ of the sentence. Each of neither and nor is tied to its corresponding item
as a COORD, and then [blue] one is tied to [pink] one as an ENUM.
POSTMODifier identifies a postposed nominal modifier, which might be either an
adjective or another noun. POSTMOD can express resultative relations (I painted the barn
red; let’s keep it a private affair), and can also serve when a modifying phrase gravitates
toward the end of the phrase due to size (he dug a hole wider than a pickup truck). There
may be intervening content in the sentence (she made it out of the haunted house alive,
where alive is a POSTMOD to she).
As with MOD, the head of POSTMOD is the noun it modifies.
Note: There’s another fringe use of POSTMOD in sentences such as the woman leapt up,
hat askew. Here hat askew creates an unusual problem since it appears syntactically to
belong to the sentence, yet is not explicitly tied in with a preposition or a verb. Use the
POSTMOD category to tie the “loose” NP to the preceding NP it describes. So hat should
tie to woman as a POSTMOD.
(Don’t confuse this with an appositive, where two NPs act as alternating names for the
same entity. See section 18.)
*PAR: I've never met anyone smarter .
%mor: pro:sub|I~aux|have adv|never part|meet&PASTP pro:indef|anyone
adj|smart-CP .
%gra: 1|4|SUBJ 2|4|AUX 3|4|JCT 4|0|ROOT 5|4|OBJ 6|5|POSTMOD 7|4|PUNCT
This is a very basic instance of POSTMOD, where smart modifies anyone but is postposed.
Part 3: Morphosyntax
*PAR: you're kind_of making me nervous .
%mor: pro:sub|you~aux|be&PRES adv|kind_of part|make-PRESP pro:obj|me
adj|nervous .
%gra: 1|4|SUBJ 2|4|AUX 3|4|JCT 4|0|ROOT 5|4|OBJ 6|5|POSTMOD 7|4|PUNCT
making me nervous exhibits a resultative relationship. me ties as an OBJ to making, while
nervous, the resultative complement, ties to me as a POSTMOD. Other verbs like get, keep,
and put can participate in similar constructions (e.g. get the crowd excited, keep it secret,
put the blazer on).
*PAR: she burned the whole boat , oars and all .
%mor: pro:sub|she v|burn-PAST art|the adj|whole n|boat cm|cm n|oar-PL conj|and pro:indef|all .
%gra: 1|2|SUBJ 2|0|ROOT 3|5|DET 4|5|MOD 5|2|OBJ 6|5|LP 7|5|POSTMOD 8|7|CONJ
9|8|COORD 10|2|PUNCT
Here we see a detached phrase, oars and all, serving to add information to boat yet not
explicitly connected to it by a verb or preposition. We’ll solve this by tying oar to boat as
a POSTMOD. and all then connect to oar using standard CONJ-COORD relations.
The trickiest question is where to tie the comma. Strictly speaking, this is not a list, but
the whole boat, oars and all still gives the impression of being one large constituent in the
sentence, so tying the comma all the way back to the root (as one would if the comma were
separating distinct clauses) is unsatisfactory. Instead treat it as you would a list, tying the
comma to the preceding boat, to preserve the constituency of the whole boat, oars and all.
COMPlement identifies a finite clausal complement to a verb, such as a subordinate clause
introduced by a verb like say, know, or think (they know we’re here). The main verb of this
subordinate clause, being its topmost element, is the item that should be classed as COMP.
The head of COMP will then be the verb for which the clause is acting as a complement.
LINK identifies a word in a number of categories that can introduce a subordinate clause,
such as complementizers (that in I heard that you’d given it up), relativizers (who in the
girl who told me about the concert), and subordinate conjunctions (unless in we’re fine
unless the power goes out).
These do not necessarily have to be subordinate clauses, though; an English speaker may
easily utter the sentence unless the power goes out on its own, in which case unless is still
Part 3: Morphosyntax
classed as a LINK. This is especially common with sentences beginning with and, but, or,
The head of LINK is the main verb of its clause.
*PAR: I can't believe you helped him .
%mor: pro:sub|I mod|can~neg|not v|believe pro:subj|you v|help-PAST pro:obj|him .
%gra: 1|4|SUBJ 2|4|AUX 3|2|NEG 4|0|ROOT 5|6|SUBJ 6|4|COMP 7|6|OBJ 8|4|PUNCT
*PAR: I can't believe that you helped him .
%mor: pro:sub|I mod|can~neg|not v|believe rel|that pro:subj|you v|help-PAST pro:obj|him .
%gra: 1|4|SUBJ 2|4|AUX 3|2|NEG 4|0|ROOT 5|7|LINK 6|7|SUBJ 7|4|COMP 8|7|OBJ 9|4|PUNCT
Here are two minimally contrasting examples of COMP. believe here takes a COMP rather
than an OBJ because its non-subject argument is an entire clause. The main verb of that
complement clause, helped, is the item identified as COMP and tied to believe. From there
it behaves just as if it were the root verb of its own sentence, taking its own SUBJ and OBJ,
you and him. The only difference between examples 10a and 10b is the presence in 10b of
the complementizer that. This is a LINK which ties to the main verb of the clause it
introduces – therefore to helped.
*PAR: not that the voters pay attention to those nuances .
%mor: neg|not rel|that art|the n|voter-PL v|pay n|attention prep|to pro:dem|those n|nuance-PL .
%gra: 1|2|NEG 2|5|LINK 3|4|DET 4|5|SUBJ 5|0|ROOT 6|5|OBJ 7|5|JCT 8|9|DET 9|7|POBJ
that as a LINK reappears, but here, it’s not embedded in any kind of subordinate clause.
Rather, it applies over the entire sentence. That’s not a problem at all, however; it still ties
as a LINK to the main verb of the clause it introduces, which is of course the main verb of
the sentence, pay.
The other interesting wrinkle here is the presence of not. It is, in fact, negating the LINK
that itself, and therefore ties to it. There are many English formulas in which not negates a
LINK (other examples: not if I have anything to say about it; I’m doing this not because
you asked, but because I want to; it’ll happen not when you’re expecting it, but when you’ll
Part 3: Morphosyntax
be surprised by it). The NEG ties to the LINK in each of these cases.
11.10 XCOMP and INF
XCOMPlement identifies a non-finite clausal complement to a verb. For example, this
might be an infinitive verb phrase introduced by try, begin, want, etc. (just try to act
natural). In addition to infinitive verbs, this can also include other non-tensed verbs like
the present participle (just try acting natural).
Just as with COMP, the main verb of this clause is classed as XCOMP and its head is the
verb for which it acts as a complement.
INFinitive identifies the infinitival particle to. The head of INF is the verb which it
*PAR: and now the cat wants to eat a can of tuna .
%mor: conj|and adv|now art|the n|cat v|want-3S inf|to v|eat art|a n|can prep|of n|tuna .
%gra: 1|5|LINK 2|5|JCT 3|4|DET 4|5|SUBJ 5|0|ROOT 6|7|INF 7|5|XCOMP 8|9|DET 9|7|OBJ
10|9|NJCT 11|10|POBJ 12|5|PUNCT
There are several things to note here. As with the relation between believe and helped in
examples 10a & 10b, the relation between want and eat is that of a verb taking a clausal
complement. However, in this case the verb acting as the complement is non-finite, so eat
ties to want as an XCOMP rather than a COMP. We also see the infinitival particle to, tied
to eat as an INF.
Finally, opening with and is another of the ways a sentence might legally begin with a
LINK. Remember that and is not tagged as a CONJ here. (It’s not conjoining two
equivalent constituents within one sentence.)
11.11 QUANT and PQ
QUANTifier identifies a nominal quantifier, such as many or several. Numbers that
quantify nouns also fall into this category (three foxes), as do distributive terms like each
and any. However, in all these cases, the QUANT must be modifying a noun phrase; if
there is no noun phrase being modified, then the quantifier is presumably acting as a
pronoun and should be classified as such. (For example, in some say he’s a traitor, some
is the SUBJ, not a QUANT.)
Like MODs, QUANTs do not stack when they quantify the same noun phrase, nor do
they interact with MODs or a DET. All tie individually to the head noun.
The head of QUANT is the noun it modifies.
PostQuantifier = PQ identifies a quantifier that follows rather than precedes its noun
phrase. This mostly occurs with both (we both tried it) and all (they all gave up). As with
POSTMOD, it’s acceptable for some sentential content to intervene between the quantified
Part 3: Morphosyntax
noun and the PQ.
The head of PQ is the noun it modifies.
*PAR: could you grab me two yellow onions and some carrots ?
%mor: mod|could pro:sub|you v|grab pro:obj|me det:num|two adj|yellow n|onion-PL conj|and
qn|some n|carrot-PL ?
%gra: 1|3|AUX 2|3|SUBJ 3|0|ROOT 4|3|OBJ2 5|7|QUANT 6|7|MOD 7|3|OBJ 8|7|CONJ
9|10|QUANT 10|8|COORD 11|3|PUNCT
Two examples of QUANTs modifying noun phrases: two and some. Just like DET,
QUANT doesn’t interact with MOD at all, but instead ties separately to the noun it
quantifies, while MOD does the same. So two ties to onions here, not yellow.
*PAR: you should really both try to apologize .
%mor: pro:sub|you mod|should adv|real&dadj-LY qn|both v|try inf|to v|apologize .
%gra: 1|5|SUBJ 2|5|AUX 3|5|JCT 4|1|PQ 5|0|ROOT 6|7|INF 7|5|XCOMP 8|5|PUNCT
It is not a problem that should and really intervene between you and both here. both still
ties to you as a PQ, since that’s the term it serves to quantify.
*PAR: Maisie picked up some more new skills .
%mor: n:prop|Maisie v|pick-PAST prep|up qn|some qn|more adj|new n|skill-PL .
%gra: 1|2|SUBJ 2|0|ROOT 3|2|JCT 4|7|QUANT 5|7|QUANT 6|7|MOD 7|3|POBJ 8|2|PUNCT
Not only do we see skills both quantified and modified here, there are two QUANTs, some
and more. A more subtle linguistic analysis might possibly suggest that some should tie to
more, but for consistency with MOD and to make the process easier for GRASP to
generalize, we tie both QUANTs individually to skills, while new also ties separately to it.
Part 3: Morphosyntax
Clausal SUBJect = CSUBJ, Clausal OBJect = COBJ, Clausal Prepositional OBJect =
CPOBJ, and Clausal PREDicate = CPRED are variations on the categories SUBJ, OBJ,
POBJ, and PRED. In each of these cases, the appropriate syntactic role is being performed
by an entire verbal clause (which could be either finite or non-finite) rather than a noun
The main verb of that clause, being the topmost element, is the one that will be identified
as CSUBJ, etc. From there, the clause ties to the rest of the sentence exactly as its nonclausal equivalent would. So the head of CSUBJ, for example, is the verb for which that
clause acts as the subject.
*PAR: Leroy getting killed off was a terrible ending.
%mor: n:prop|Leroy part|get-PRESP part|kill-PASTP prep|off cop|be&PAST&13S art|a adj|terrible
n|ending .
%gra: 1|3|SUBJ 2|3|AUX 3|5|CSUBJ 4|3|JCT 5|0|ROOT 6|8|DET 7|8|MOD 8|5|PRED 9|5|PUNCT
The entire clause Leroy getting killed off is the subject which is ascribed the quality of
being a terrible ending. Thus the main verb of that clause, killed, is tied as a CSUBJ to
*PAR: I want whoever keeps calling me to stop .
%mor: pro:sub|I v|want pro:wh|whoever v|keep-3S part|call-PRESP pro:obj|me inf|to v|stop .
%gra: 1|2|SUBJ 2|0|ROOT 3|4|SUBJ 4|2|COBJ 5|4|XCOMP 6|5|OBJ 7|8|INF 8|2|XCOMP
whoever keeps calling me is a clause here acting as the object of want. So the main verb of
this clause, keeps, ties to want as a COBJ.
*PAR: I'm not clear on why she did that .
%mor: pro:sub|I~cop|be&1S neg|not adj|clear prep|on adv:wh|why pro:sub|she v|do-PAST
pro:dem|that .
%gra: 1|2|SUBJ 2|0|ROOT 3|2|NEG 4|2|PRED 5|4|JCT 6|8|LINK 7|8|SUBJ 8|5|CPOBJ 9|8|OBJ
Part 3: Morphosyntax
The clause why she did that acts as the object of the preposition on; thus its main verb did
ties to on as a CPOBJ.
*PAR: the trick is getting the coin to land just right .
%mor: art|the n|trick cop|be&3S part|get-PRESP art|the n|coin inf|to v|land adv|just adv|right .
%gra: 1|2|DET 2|3|SUBJ 3|0|ROOT 4|3|CPRED 5|6|DET 6|4|OBJ 7|8|INF 8|4|XCOMP 9|10|JCT
10|8|JCT 11|3|PUNCT
getting the coin to land just right is the predicate to the main verb be; thus its main verb
getting ties to be as a CPRED.
11.13 CJCT and XJCT
Clausal conJunCT = CJCT identifies a finite clause that functions as an adjunct to a verb,
adjective, or adverb. For example, in the sentence you’ll do it because I said so, the phrase
because I said so is a clausal conjunct on the topmost verb do.
As with the clausal categories introduced in the previous section, the main verb within
the adjoining clause (said in the above example) is the item that should be tagged CJCT,
and its head is the main verb of the clause to which it adjoins (do).
XadJunCT = XJCT identifies a non-finite clause (that is, generally, one headed by an
infinitive verb or participle) that otherwise behaves like a CJCT, as in to build a ship, first
gather some wood.
Again, the main verb in the adjoining clause (build in the example) is tagged as XJCT
and its head is the main verb of the clause to which it adjoins (gather).
Note: In a sentence like I popped the cork and champagne sprayed out, we treat the second
clause (and champagne sprayed out) as a CJCT to the main clause (I popped the cork). So
we tie sprayed to popped as a CJCT, and treat and as a LINK tied to sprayed.
The alternative would be using a CONJ-COORD relation between the two verbs (you
could tie and to popped as a CONJ, then sprayed to and as a COORD) but this is
dispreferred because the two clauses have different subjects.
On the other hand, in a sentence like I popped the cork and poured everyone a round, it’s
acceptable to tie popped – and – poured together in a CONJ-COORD relation, since here
both verbs have the same subject and can be treated as coordinated verb phrases.
Part 3: Morphosyntax
*PAR: try not to move until we've brought you to the infirmary .
%mor: v|try neg|not inf|to v|move conj|until pro:sub|we~aux|have part|brought-PASTP pro:obj|you
prep|to art|the n|infirmary .
%gra: 1|0|ROOT 2|4|NEG 3|4|INF 4|1|XCOMP 5|8|LINK 6|8|SUBJ 7|8|AUX 8|1|CJCT 9|8|OBJ
10|8|JCT 11|12|DET 12|10|POBJ 13|1|PUNCT
until we have brought you to the infirmary is a finite clausal adjunct, and so its main verb
brought ties to try as a CJCT. until functions as a LINK introducing the adjoining clause.
*PAR: we spent the whole day visiting museums and galleries .
%mor: pro:sub|we v|spend-PAST art|the adj|whole n|day part|visit-PRESP n|museum-PL
conj|and n|gallery-PL .
%gra: 1|2|SUBJ 2|0|ROOT 3|5|DET 4|5|MOD 5|2|OBJ 6|2|XJCT 7|6|OBJ 8|7|CONJ 9|8|COORD
visiting museums and galleries is a non-finite clausal adjunct whose main verb ties to spend
as an XJCT.
11.14 CMOD and XMOD
Clausal MODifier = CMOD identifies a finite clause that modifies a noun, as in that boy
Tricia took to prom. CMOD also sometimes modifies an adjective or adverb, if that
adjective/adverb takes a complement (we’re happy that you made it) or participates in a
resultative expression (so quietly I couldn’t hear them).
CMOD functions much like CJCT and XJCT: the main verb within the modifier clause
is the item tagged CMOD, and its head is the noun, adjective, or adverb it modifies.
XMODifier identifies a non-finite clause that otherwise behaves like a CMOD, as in the
person running the store; a good day to take a walk; she’s eager to teach them.
Once again, the main verb within the modifying clause is the one tagged XMOD, and its
head is the noun, adjective, or adverb it modifies.
*PAR: I'm throwing away the shoes I wore for the race .
%mor: pro:sub|I~aux|be&1S part|throw-PRESP prep|away art|the n|shoe-PL pro:sub|I v|wear-
Part 3: Morphosyntax
PAST prep|for art|the n|race .
%gra: 1|3|SUBJ 2|3|AUX 3|0|ROOT 4|3|JCT 5|6|DET 6|4|POBJ 7|8|SUBJ 8|6|CMOD 9|8|JCT
10|11|DET 11|9|POBJ 12|3|PUNCT
Notice that here the clausal adjunct I wore for the race does not modify throwing, but rather
shoes – this clause specifies a particular pair of the speaker’s shoes. Therefore wore ties to
shoes as a CMOD rather than tying to throwing as a CJCT.
*PAR: Isabelle is the one to ask about computer problems .
%mor: n:prop|Isabelle cop|be&3S art|the n|one inf|to v|ask prep|about n|computer n|problem-PL .
%gra: 1|2|SUBJ 2|0|ROOT 3|4|DET 4|2|PRED 5|6|INF 6|4|XMOD 7|6|JCT 8|9|MOD 9|7|POBJ
Again, here, the non-finite clausal adjunct to ask about computer problems does not modify
the main verb is. Instead it specifies the predicate noun, one, and so ties to it as an XMOD.
*PAR: they were so hungry that they couldn't sleep .
%mor: pro:sub|they cop|be&PAST adv|so adj|hungry rel|that pro:sub|they mod|could~neg|not
v|sleep .
%gra: 1|2|SUBJ 2|0|ROOT 3|4|JCT 4|2|PRED 5|9|LINK 6|9|SUBJ 7|9|AUX 8|7|NEG 9|4|CMOD
A typical so X that Y formation. There’s a conceivable argument that the subordinate clause
should simply tie to be as a CJCT, but one senses that so hungry that they couldn’t sleep is
a single constituent unto itself. That is how it’s captured here, headed up by hungry, with
the subordinate clause acting as a CMOD to that adjective.
BEGinning identifies a clause-external element occurring at the beginning of an utterance.
This might be, for example, an interjection like hey or okay, or a vocative like the
addressee’s name.
The head of BEG is 0, the LeftWall.
BEGinning Punctuation = BEGP identifies the double dagger ‡ which is employed to
partition off a BEG element from the rest of the sentence for the purposes of MOR. The
double dagger should always follow a BEG.
Part 3: Morphosyntax
The head of BEGP is the BEG it follows.
END identifies a clause-external element occuring at the end of an utterance. This includes
interjections, vocatives, and other elements similar to those that can act as BEGs. However,
it doesn’t include tag questions like …don’t you? (See section 17.)
Notably, whereas the head of BEG is 0, the head of END is the ROOT.
END Punctuation = ENDP identifies the double comma „ which is employed to partition
off an END element much like the double dagger. The double comma should always
precede an END or a TAG (again see section 17).
The head of ENDP is the END it precedes.
*PAR: okay ‡ here's the floor she's trying to rent out .
%mor: co|okay beg|beg pro:exist|here~cop|be&3S art|the n|floor pro:sub|she~aux|be&3S part|tryPRESP inf|to v|rent prep|out .
%gra: 1|0|BEG 2|1|BEGP 3|4|JCT 4|0|ROOT 5|6|DET 6|4|PRED 7|9|SUBJ 8|9|AUX 9|6|CMOD
10|11|INF 11|9|XCOMP 12|11|JCT 13|4|PUNCT
A simple BEG construction. The BEG okay ties directly to 0, while the BEGP ties to okay.
*PAR: you'll probably want to keep an eye on that „ Hugh .
%mor: pro:sub|you~mod|will adv|probable&dadj-LY v|want inf|to v|keep art|an n|eye prep|on
pro:dem|that end|end n:prop|Hugh .
%gra: 1|4|SUBJ 2|4|AUX 3|4|JCT 4|0|ROOT 5|6|INF 6|4|XCOMP 7|8|DET 8|6|OBJ 9|8|NJCT
10|9|POBJ 11|12|ENDP 12|4|END 13|4|PUNCT
On the other hand, the END here, Hugh, ties to the ROOT want (then the ENDP ties to
11.16 COM and TAG
COMmunicator identifies any clause-external element occuring medially rather than
initially or finally. Depending on the transcriber, such an element might be partitioned off
with commas (I walked into the courtroom, right, and here’s this guy telling me…), or it
might not be partitioned at all (I know you don’t like it Li but it’s the best solution we’ve
Like END, and unlike BEG, the head of COM is the ROOT.
TAG identifies the verb of a tag question at the end of a sentence (…doesn’t it?). As
mentioned in the previous section, this is not considered an END, but has a category of its
Part 3: Morphosyntax
own since it takes a more complex set of syntactic interactions. However, as with END,
there may be a double comma preceding the tag verb; if so, tie it to the tag verb as an
The head of TAG, like END, is the ROOT.
*PAR: I think it's possible , sure , but who can say ?
%mor: pro:sub|I v|think pro|it~cop|be&3S adj|possible cm|cm co|sure cm|cm conj|but pro:wh|who
mod|can v|say ?
%gra: 1|2|SUBJ 2|0|ROOT 3|4|SUBJ 4|2|COMP 5|4|PRED 6|2|LP 7|2|COM 8|2|LP 9|12|LINK
10|12|SUBJ 11|12|AUX 12|2|CJCT 13|2|PUNCT
A communicator, sure, intervenes between the two clauses of this sentence. So it’s marked
as a COM and tied to the ROOT think. Both of the commas used to partition it off are also
tied to the ROOT (since they’re not delineating items in a list, but rather entire clauses).
*PAR: you've toured just about everywhere „ haven't you ?
%mor: pro:sub|you~aux|have part|tour-PASTP adv|just adv|about adv|everywhere end|end
aux|have~neg|not pro:sub|you ?
%gra: 1|3|SUBJ 2|3|AUX 3|0|ROOT 4|5|JCT 5|6|JCT 6|3|JCT 7|8|ENDP 8|3|TAG 9|8|NEG
10|8|SUBJ 11|3|PUNCT
TAG in action. The tag verb, have, ties to the ROOT toured. That tag verb takes its own
NEG not and even its own SUBJ you, so it’s more complex than an END. However, just
as if it were an END, the double comma ties to it as an ENDP.
Also worth calling to attention is the fact that the PUNCT ? still ties to the ROOT, even
though it could be more properly understood as tying to the tag verb. For convenience, we
want to tie PUNCT consistently to the ROOT in all cases.
11.17 SRL, APP
SeRiaL identifies serial verbs, such as come play and go see. In each of these pairs, the
second verb is treated as the head and the first verb is treated as the dependent; thus the
first verb is tagged SRL and its head is the second verb.
Generally serial verb pairs are both bare verbs, but there’s at least one exception: get in
the sense of get going or another get X-ing. Notice that the second verb doesn’t seem to
relate to get as an object or complement, so the function of get appears more serial.
(Contrast the more transparent start packing, try spinning, etc., where the second verb
Part 3: Morphosyntax
would tie as an XCOMP to the first verb.)
APPositive identifies an appositive noun or phrase, as in Giles the baker. The head of APP
is the noun it corresponds to. However, which of the two apposed terms should be
considered the head and which the APP will depend on the structure of the sentence.
In most straightforward constructions (as in the one above), the first term is the head and
the following term is the APP (so baker will tie to Giles as an APP). On the other hand, a
sentence might be structured such that the second term – especially if it’s a pronoun – will
read more like the SUBJ (etc.) and the first term will read more like the APP. In the
sentence myself, I prefer provolone, treat I as the SUBJ and tie myself to it as an APP.
Remember the essential quality of APP: both noun phrases denote the same entity, and
serve as two different ways of naming it. That makes APP unlike our hat askew example
from back in section 9, which is why APP is inappropriate there, and POSTMOD is used
*PAR: go check whether the rice is ready .
%mor: v|go v|check conj|whether art|the n|rice cop|be&3S adj|ready .
%gra: 1|2|SRL 2|0|ROOT 3|6|LINK 4|5|DET 5|6|SUBJ 6|2|COMP 7|6|PRED 8|2|PUNCT
A straightforward example of a serial verb. go ties to check as a SRL.
*PAR: he told his boss , Masha , he was switching to a new job .
%mor: pro:sub|he v|tell-PAST pro:poss:det|his n|boss cm|cm n:prop|Masha cm|cm pro:sub|he
aux|be&PAST&13S part|switch-PRESP prep|to art|a adj|new n|job .
%gra: 1|2|SUBJ 2|0|ROOT 3|4|MOD 4|2|OBJ 5|4|LP 6|4|APP 7|6|LP 8|10|SUBJ 9|10|AUX
10|2|COMP 11|10|JCT 12|14|DET 13|14|MOD 14|11|NJCT 15|2|PUNCT
Here Masha ties to boss as an APP. Notice that the transcriber put commas here to partition
Masha from the rest of the sentence. This is optional – it makes no difference in how Masha
ties to boss – but take note of how the commas are treated, each one tied to the word
preceding. Essentially the commas behave exactly as if this were a list of items conjoined
with ENUM (so they tie directly to those items, rather than the ROOT).
11.18 NAME, DATE
NAME identifies relations within a string of elements that are functioning together to act
as one name, like Deandre Clarks and South Carolina. According to modern transcription
protocol, these should generally be conjoined with underscores to form single syntactic
Part 3: Morphosyntax
units (Deandre_Clarks and South_Carolina), in which case they don’t pose a problem for
GRASP. However, if the two or more elements of a name aren’t or cannot be conjoined
(for instance, if there’s intervening material: Deandre &uh Clarks), the NAME tag will be
Use NAME to tie every non-final element in the name to the final element. Thus in South
Carolina, South will tie to Carolina as a NAME.
DATE identifies relations within a string of elements that are functioning together to act
as a date, like July twentieth nineteen ninety. Unlike proper names, these are never
conjoined with underscores to form single syntactic units, since the combination of
numbers would be endless. Thus DATE is required in all cases.
For consistency with NAME, tie every non-final element in the date to the final element.
Thus in July twentieth nineteen ninety, all elements but ninety will tie to it as DATEs.
*PAR: June sixteenth , nineteen oh [: zero] four was just another day for Leopold Bloom .
%mor: n:prop|June adj|sixteenth cm|cm det:num|nineteen det:num|zero det:num|four
cop|be&PAST&13S adv|just qn|another n|day prep|for n:prop|Leopold n:prop|Bloom .
%gra: 1|6|DATE 2|6|DATE 3|2|LP 4|6|DATE 5|6|DATE 6|7|SUBJ 7|0|ROOT 8|9|JCT 9|10|QUANT
10|7|PRED 11|10|NJCT 12|13|NAME 13|11|POBJ 14|7|PUNCT
Examples of both DATE and NAME. The interaction of NAME is simple: Leopold ties to
Bloom. (Again, if the transcriber were better acquainted with modern practice, he or she
would presumably have written the name as Leopold_Bloom, thus obviating the need for
NAME, but let’s consider the non-optimal case.)
The date is slightly more dense: each of the non-final elements of the date ties to the final
one, four, as a DATE, but there’s also an intervening comma after sixteenth, which, once
again, ties to the item it follows, as if this date were a list of enumerated items.
INCROOT is a catch-all category for any word that must substitute in the role of ROOT
when no proper ROOT is available. This may be because the sentence lacks a verb
altogether, or because any verb present is not top-level (say, it’s part of a clause depending
on another phrase). When no ROOT is available, choose the topmost word in the sentence
and name it the INCROOT.
The head of INCROOT is 0.
OMission is another catch-all for any determiner, modifier, or other dependent word which
is left headless due to an omission, such as the speaker trailing off or being interrupted and
hence leaving their sentence unfinished.
The head of OM is whatever the head of its head would have been. (So the head of OM
is the ROOT, for instance, if what’s been omitted was the OBJ of the root verb.)
Part 3: Morphosyntax
*PAR: the one with the stripes ?
%mor: art|the n|one prep|with art|the n|stripe-PL ?
%gra: 1|2|DET 2|0|INCROOT 3|2|NJCT 4|5|DET 5|3|POBJ 6|2|PUNCT
Here is an example of INCROOT in action. Since there’s no verb in this sentence, there is
no proper ROOT, and we must assign the label of INCROOT to another element in the
sentence. The topmost existing element is one, upon which the determiner the depends, as
well as the entire NJCT phrase that follows. Thus it is the logical choice to be identified as
the INCROOT. Naturally, PUNCT will now tie to one as if it were the ROOT.
*PAR: Eloise wrote the +...
%mor: n:prop|Eloise v|write-PAST art|the +...
%gra: 1|2|SUBJ 2|0|ROOT 3|2|OM 4|2|PUNCT
A simple case of omission. Lacking any other context, it’s safe to assume that the was
going to introduce the noun phrase functioning as the OBJ of wrote. Therefore we’ll tie the
with the label OM to wrote – one branch higher up the tree, so to speak, than what it was
meant to attach to if the sentence had been completed.
Part 3: Morphosyntax
12 GRs for other languages
Although many of the basic GRs for English can be found in other languages, there are
often relations in these other languages that are not found in English.
12.1 Spanish
The GRs needed for Spanish are close to those for English. The distinction between X and
C in the XJCT/CJCT and XMOD/CMOD relations are irrelevant to Spanish because the
finite/nonfinite status of the verb is clearly reflected in its morphology. So, the XJCT and
XMOD relations can be dropped. The two additional relations that are needed are
APREP for the a-personal. This relation can either mark the direct object, when no
OBJ is present or the indirect object when there is an OBJ relation. The APREP
attaches to the verb and the object of the preposition attaches to the “a” with the
POBJ relation.
INF for the preposition introducing an infinitive. For example, in lo hizo para salir,
the preposition para relates to the infinitive salir with the INF relation. The
infinitive then relates to the main verb with a COMP relation.
12.2 Chinese
The GRs needed for Chinese are not too very different from those needed for English. The
major differences involve these areas:
1. Chinese uses all the basic GRs that are also found in English with these exceptions:
TAG, DET, and INF.
2. Also, Chinese does not have a finite, non-finite distinction on verbs. Somewhat
arbitrarily, this makes the “X” relations of English irrelevant and only CSUBJ,
COMP, CMOD, CPRED, and CJCT are needed. Also, the main verb is the head of
the clause, not the auxiliary.
3. Chinese makes extensive use of topics and sentence final particles. In the TOP
relation, the head/root can be a nominal or adjective, as well as a verb.
4. Chinese has a variety of verb-verb constructions, beyond simple serial verbs. For
Chinese, the head of SRL is the first verb, not the second.
5. The Chinese possessive construction has the head in final position.
6. Chinese can use classifier phrases such as yi1 bian4 as JCT.
7. Chinese often combines clauses without using any conjunction to mark subordination
or coordination. In this case, it is best to transcribe the two clauses as separate
sentences. To mark the missing subordination relation, just add this postcode to the
end of the first sentence: [+ sub]. This mark does not necessarily imply which clause
is subordinate; it just notes that the two clauses are related, although the relation is
not marked with a conjunction.
8. The CJCT relation can also extend to mei2 you3 because they are both words
neg|mei2=not v|you3=have v|jiang3=speak v:resc|wan2=finish .
1|2|NEG 2|3|CJCT 3|0|ROOT 4|3|VR 5|3|PUNCT
9. The CJCT relation is also used in the many cases where there are no subordinating
conjunctions, as in this example:
n|qing1wa1 adv|yi1 v|kan4 adv|jiu4 neg|mei2 sfp|le
Part 3: Morphosyntax
1|3|SUBJ 2|3|JCT 3|5|CJCT 4|5|JCT 5|0|INCROOT 6|5|SFP
10. Clefts can be coded using the CPRED relations as in this example:
co|lai2 n:relat|jie3&DIM adv|jiu4 v:cop|shi4 adv|zhe4yang4 v|jiang3 sfp|de
1|0|COM 2|4|SUBJ 3|4|JCT 4|0|ROOT 5|6|JCT 6|4|CPRED 7|4|SFP 8|4|PUNCT
11. The JCT relation for Chinese extends also to the complements of directional verbs, as
in this example:
v|pao3=run v:dirc|jin4=enter n|ping2=bottle post|li3=inside
1|0|ROOT 2|1|VD 3|4|POSTO 4|1|JCT
Note that the JCT is attached to the second of the two serial verbs
The additional relations needed for Chinese are:
1. POSSession = POSS is the relation that holds between the linker “de” and the
preceding possessor noun or pronoun which then functions as the head for further
attachment to the thing possessed through the MOD relation.
2. Chinese has no articles and uses classifiers (CLASS) to mark quantification.
3. VR is the relation between the initial verb as head and the following resultative verb
as dependent.
4. VD is the relations between the initial verb in a serial verb construction as head and
the following directional verb as dependent.
5. SFP is a relation between the sentence final particle and “0”.
6. PTOP is a postposed topic, as in this example:
pro|zhe4=this class|ge4 v:cop|shi4=is pro:wh|shen2me=what pro|zhe4=this class|ge4
1|2|DET 2|3|SUBJ 3|0|ROOT 4|3|OBJ 5|6|DET 6|3|SUBJ
7. PostpositionalObject = POSTO is the relation between a postposition, which is the
head and the dependent noun. This relation is marked on the noun, as in this example:
ta pao dao jiaoshu limian qu
1|2|SUBJ 2|0|ROOT 3|2|JCT 4|5|POSTO 5|3|JCT 6|2|SRL
8. PrepositionalObject = PREPO is the relation between a preposition, which is the head
and the following dependent noun, as in this example, where the relation is coded on
the noun:
ta shi cong Beijing lai de
1|2|SUBJ 2|0|ROOT 3|5|JCT 4|3|PREPO 5|2|CJCT 6|5|LINK
9. This relation can also hold between a preposition and a postposition, as in this
v|yang3=maintain prep|zai4=at pro:wh|shen2me=what post|li3mian4=inside ?
1|0|ROOT 2|1|JCT 3|4|POSTO 4|2|PREPO 5|1|PUNCT
12.3 Japanese
The GRs needed for Japanese are more differentiated than those needed for English. The
major differences involve these areas:
1. Japanese is left-branched with the head following the modifier, adjunct or
2. Japanese uses optional case particles to identify case relations. The head noun is coded
Part 3: Morphosyntax
as SUBJ or OBJ, resp. The case particles are coded as CASP, and constitute the head
of the argument. Free adjuncts are coded as JCT, and the corresponding postparticles
Ken ga Tookyoo kara kita
Ken SUBJ Tokyo from come-PAST “Ken came from Tokyo”
1|2|SUBJ 2|5|CASP 3|4|JCT 4|5|POSTP 5|0|ROOT
3. The root can be a tense-bearing element like a verb, a verbal adjective (ROOT), a
copula (COPROOT) or a noun (PREDROOT).
4. The root can be also a case particle (CASPROOT), a postparticle (POSTPROOT), an
attributive particle (ATTPROOT), a quotative particle (QUOTPROOT), a topic
particle (TOPPROOT), a quotative marker, a focus particle (FOCPROOT), or a
conjunctive particle (CPZRROOT). Note that these different types of roots are fully
grammatical and not elliptic fragments.
Papa no.
Dad GEN “(it's) Dad's one.”
iku kara.
go-PRES because “because (I) will go”
5. Japanese expresses topic relations (TOP); the topic particle is coded as TOPP.
6. Like Chinese, Japanese has no articles and uses classifiers and counters to mark
sankurambo sanko tabeta.
cherry 3-pieces eat-PAST “he ate 3 cherries”
1|3| OBJ 2|1|QUANT 3|0|ROOT
7. Spoken Japanese makes extensive use of sentence final particles (SFP) and sentence
modifiers (SMDR). They are depending of the preceding ROOT.
tabeta no ?
eat-PAST SFP “did you eat (it)?”
1|0|ROOT 2|1|SFP
tabeta jan.
eat-PAST SMDR “you ate it, didn't you?”
1|0|ROOT 2|1|SMDR
verbal ROOT; relation between verb and left wall: v,
adj, subsidiary verb (tense bearing element)
taberu. ‘I’ll eat it.’
COPula ROOT; copula with noun, adjectival noun, or
sentence nominalizer (no da)
koko da. ‘it’s here.’
nominal ROOT (without copula); includes adv, co,
and quant, as well as verbal nouns and adjectival nouns
in root position.
koko. ‘it’s here.’
Part 3: Morphosyntax
CPREDROOT nominal ROOT with a sentence nominalizer in root
position (ptl:snr|no=SNR)
uma no chiisai no. ‘a small horse’
CASe Particle ROOT (the case particle is the head)
dare ga? ‘who?’
TOPic Particle ROOT
kore wa? ‘and what about this one?’
FOCus Particle ROOT
kore mo? ‘this one, too?’
TOPicalization, (for convenience the root of the
sentence is considered to be the head)
kore wa yomenai. ‘I can’t read it.’
1|2|TOP 2|3|TOPP 3|0|ROOT
finite Clausal TOPic (head of the clause is ptl:snr|no)
iku no wa ii kedo [...] ‘it’s ok to go, but…’
1|2|CMOD 2|3|CTOP 3|4|TOPP 4|5|COMP 5|6|CPZR
FOCus (followed by ptl:foc; mo, shika, bakari, hodo
kore mo yonda. ‘he read this one, too.’
1|2|FOC 2|3|FOCP 3|0|ROOT
nonclausal SUBject
Jon ga tabeta. ‘John ate it.’
1|2|SUBJ 2|3|CASP 3|0|ROOT
finite Clausal SUBJect (head of the clause is ptl:snr)
taberu no ga ii. ‘it’s good to eat it.’
1|2|CMOD 2|3|CSUBJ 3|4|CASP 3|0|ROOT
accusative OBJect
hon o yonda. ‘he has read the book.’
1|2|OBJ 2|3|CASP 3|0|ROOT
finite Clausal accusative OBJect
taberu no o yameta. ‘he stopped eating.’
1|2|CMOD 2|3|COBJ 3|4|CASP 4|0|ROOT
Part 3: Morphosyntax
adJunCT (Postpositional or adverbial phrase)
gakkoo kara kaetta. ‘he came back from school.’
1|2|JCT 2|3|POSTP 3|0|ROOT
yukkuri shabetta. ‘he talked slowly.’
1|2|JCT 2|0|ROOT
finite Clausal adJunCT
ochita no de taberu. ‘I’ll eat with the one that had
fallen down.
1|2|CMOD 2|3|CJCT 3|4|POSTP 4|0|ROOT
nonfinite clause as adJunCT (tabe-reba, -tara, -te, -cha,
-tari; oishi-ku; shizuka ni)
kaeseba ii. ‘it’s ok to give it back.’
1|2|XJCT 2|0|ROOT
ComPlementiZeR (subordinating conjunctive particle;
osoi kara kaeru. ‘I’ll go home because it’s late.’
1|2|COMP 2|3|CPZR 3|0|ROOT
conjunction); head is always the root
dakara kaeru. ‘that’s why I’ll go home.’
1|2|ZCPZR 2|0|ROOT
finite clausal verb COMPlement (before ptl:conj| and
quot|to )
osoi kara kaeru. ‘I’ll go home because it’s late.’
1|2|COMP 2|3|CPZR 3|0|ROOT
QUOTation Particle after nominal or verbal phrase
kaeru to iimashita. ‘he said he would go home.’
Zero-QUOTative (sentence introducing quotative
tte iu ka […] ‘in other words’
1|2|ZQUOT 2|3|COMP 3|4|CPZR
nonclausal MODifier (of a nominal)
Papa no kutsu ga atta. ‘there are Dad’s shoes.’
1|2|MOD 2|3|CASP 3|4|SUBJ 4|5|CASP 5|0|ROOT
finite Clausal MODifier of a nominal; the dependent is
a finite verb, adjective or adj noun with copula
akai kuruma o mita. ‘he saw a red car.’
1|2|CMOD 2|3|OBJ 3|4|CASP 4|0|ROOT
nonfinite clausal MODifier of a nominal (adn|)
kore to onaji mono ga […] ‘a thing similar to this one’
1|2|JCT 2|3|POSTP 3|4|XMOD 4|5|SUBJ 5|6|CASP
COORDination, second noun is the head; (ptl:coo|)
inu to neko o katte iru. ‘he has a dog and a cat.’
1|2|COORD 2|3|COOP 3|4|OBJ 4|5|CASP 5|6|XJCT
Part 3: Morphosyntax
NP structure
nominal PREDicate before copula or QUOT
tabeta hito da. ‘he is the one who ate it.’
finite Clausal PREDicate before copula (no da)
taberu no da. ‘in fact, he’ll eat it.’
CASe Particles (ptl:case; ga, o)
hon o yonda. ‘he read the book.’
1|2|OBJ 2|3|CASP 3|0|ROOT
POSTpositional Particles (ptl:post; ni, de, kara, made,
Papa ni ageta. ‘he gave it to Dad.’
1|2|JCT 2|3|POSTP 3|0|ROOT
ATTributive Particle
Papa no kutsu (ga) ‘Dad’s shoes are…’
1|2|MOD 2|3|ATTP 3|4|SUBJ
ATTributive Particle in SUBJect position with headnoun elided
Papa no ga atta. ‘here is Dad’s one.’
1|2|MOD 2|3|ATTP-SUBJ 3|4|CASP 4|0|ROOT
ATTributive Particle in OBJect position
Papa no o mita. ‘I saw Dad’s one.’
1|2|MOD 2|3|ATTP-OBJ 3|4|CASP 4|0|ROOT
ATTributive Particle in ADJunct position
Papa no de asonda. ‘he played with Dad’s one.’
1|2|MOD 2|3|ATTP-JCT 3|4|POSTP 4|0|ROOT
ATTributive Particle in predicate position
Papa no da. ‘it’s Dad’s one.’
ATTributive Particle in TOPic position
Papa no wa agenai. ‘I won’t give you Dad’s one.’
1|2|MOD 2|3|ATTP-TOP 3|4|TOPP 4|0|ROOT
ATTributive Particle in FOCus position
Papa no mo agenai. ‘I also won’t give you Dad’s one.’
1|2|MOD 2|3|ATTP-FOC 3|4|TOPP 4|0|ROOT
TOPic Particle (ptl:top; wa)
kore wa yomenai. ‘I can’t read this.’
1|2|TOP 2|3|TOPP 3|0|ROOT
FOCus Particle (ptl:foc; mo, shika, bakari, hodo etc.)
kore mo yonda. ‘I read this one, too.’
1|2|FOC 2|3|FOCP 3|0|ROOT
COOrdination Particles (ptl:coo; to, ya etc.)
inu to neko ga […] ‘dogs and cats are…’
1|2|COORD 2|3|COOP 3|4|SUBJ 4|5|CASP
QUANTifier (incl. classifiers and counters)
banana sambon tabeta. ‘he ate three bananas.’
Part 3: Morphosyntax
1|3|OBJ 2|3|QUANT 3|0|ROOT
ENUMeration, without coordinating particle
ichi ni sanko da. ‘there are 1, 2, 3 of them.’
1|2|ENUM 2|3|ENUM 3|4|PRED 4|0|ROOT
string of proper NAMEs, second name is the head
Kameda Taishoo ga kita. ‘Taishoo Kameda arrived.’
1|2|NAME 2|3|SUBJ 3|4|CASP 4|0|ROOT
string of DATEs, last element (day) is the head
rokugatsu tsuitachi ni kita. ‘he came on June 1st.’
1|2|DATE 2|3|JCT 3|4|POSTP 4|0|ROOT
sentence final Sentence MoDifieR (smod| ; mitai, jan,
rashii etc); for convenience, the tense bearing verb is
considered to be the head
kaetta mitai. ‘it seems he went home.’
1|0|ROOT 2|1|SMDR
Sentence Final Particle (including the use after
arguments and adjunct)
kuru ne. ‘he’ll come, won’t he?’
1|0|ROOT 2|1|SFP
COMmunicator; (co:i| co:g|) including isolated final
particles, sentence modalizers and onomatopoeias;
head is always set to 0
anoo tabeta. ‘err..I ate it.’
1|0|COM 2|0|ROOT
VOCative ; head is always set to 0
Taishoo ‡ aka. ‘Taishoo, it’s red.’
sentence boundary (sentence ends; .!? etc.); the root is
the head
iku . ‘I’ll go.’
1|0|ROOT 2|1|PUNCT
Right Dislocation boundary (dloc|„=DISLOC);
dislocation follows; the root is the head
mita „ fuusen ? ‘the balloon, did you see it?’
1|0|ROOT 2|1|RDP 3|1|OBJ 4|1|PUNCT
VOCative marker (voc|‡=VOC ); head is the
preceding vocative
Taishoo ‡ mite ! ‘Taishoo, look!’
1|0|VOC 2|1|VOCP 3|0|ROOT 4|3|PUNCT
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF