Lecture 7 . ...... Syntactic Formalisms for Parsing Natural Languages Aleš Horák, Miloš Jakubíček, Vojtěch Kovář (based on slides by Juyeon Kang) ia161@nlp.fi.muni.cz Autumn 2013 IA161 Syntactic Formalisms for Parsing Natural Languages 1 / 64 Lecture 7 . ...... Parsing with HPSG IA161 Syntactic Formalisms for Parsing Natural Languages 2 / 64 Lecture 7 Overview on syntactic formalisms Unification based grammars : HPSG, LFG, TAG, UCG... Dependency based grammars : Tesnière model; Meaning-Text of Mel’čuk... IA161 Syntactic Formalisms for Parsing Natural Languages 3 / 64 Lecture 7 Heritage of HPSG GPSG – Generalized Phrase-Structure Grammar (Gerald Gazdar) linear order/hierarchy order feature structure for representation of information LFG Lexicon contains Lexical rules CG Subcategorization IA161 Syntactic Formalisms for Parsing Natural Languages 4 / 64 Lecture 7 Key points of HPSG Monostratal theory without derivation Sharing a given information without movement and transformation One representation for different levels of analysis : phonology, syntax, semantic Constraint-based analysis Unification of given information Computational formalism IA161 Syntactic Formalisms for Parsing Natural Languages 5 / 64 Lecture 7 Syntactic representation in HPSG Typed feature structure consists of a couple “attribute/value” the types are organized into a hierarchy ex: sign>phrase, case>nominative feature structure is a directed acyclic graph (DAG), with arcs representing features going between values IA161 Syntactic Formalisms for Parsing Natural Languages 6 / 64 Lecture 7 Features Basic element of structure in HPSG Should be appropriate to a type Most frequently used features PHON SYNSEM LOC/NON-LOC CAT CONTEXT CONTENT HEAD SUJ COMPS S-ARG IA161 Syntactic Formalisms for Parsing Natural Languages 7 / 64 Lecture 7 Types Types are attributed to features -> typed features sign synsem head phrase content Index .... Each of these feature values is itself a complex object: The type sign has the features PHON and SYNSEM appropriate for it The feature SYNSEM has a value of type synsem This type itself has relevant features (LOCAL and NONLOCAL) IA161 Syntactic Formalisms for Parsing Natural Languages 8 / 64 Lecture 7 Types sign is the basic type in HPSG used to describe lexical items (of type word) and phrases (of type phrase). All signs carry the following two features: PHON encodes the phonological representation of the sign SYNSEM syntax and semantics sign   PHON list(phon-string) SYNSEM synsem   IA161 Syntactic Formalisms for Parsing Natural Languages 9 / 64 Lecture 7 Types In attribute-value matrix (AVM) form, here is the skeleton of an object:               sign PHON list(PHON) SYNSEM      synsem LOCAL local NON-LOCAL non-local      DTRS list(SIGN)               IA161 Syntactic Formalisms for Parsing Natural Languages 10 / 64 Lecture 7 Structure of signs in HPSG synsem introduces the features LOCAL and NONLOCAL local introduces CATEGORY (CAT), CONTENT (CONT) and CONTEXT(CONX) non-local will be discussed in connection with unbounded dependencies category includes the syntactic category and the grammatical argument of the word/phrase IA161 Syntactic Formalisms for Parsing Natural Languages 11 / 64 Lecture 7 Description of an object in HPSG: lexical sign and phrasal sign sing [ PHON list(phon-string) SYNSEM synsem ] word phrase [ DTRS constituent-struc ] synsem   LOCAL local NON-LOCAL non-local   local      CATEGORY category CONTENT content CONTEXT context      category      HEAD head VAL ... ... ...      IA161 Syntactic Formalisms for Parsing Natural Languages 12 / 64 Lecture 7 CATEGORY CATEGORY encode the sign’s syntactic category Given via the feature [HEAD head], where head is the supertype for noun, verb, adjective, preposition, determiner, marker; each of these types selects a particular set of head features Given via the feature [VALENCE ...], possible to combine the signs with the other signs to a larger phrases     SYNSEM|LOC|CAT|VALENCE valence     SUBJECT list(synsem) SPECIFIER list(synsem) COMPLEMENTS list(synsem)         IA161 Syntactic Formalisms for Parsing Natural Languages 13 / 64 Lecture 7 Sub-categorization of head type vform finite infinitive base gerund present-part. past-part. passive-part. case nominative accusative pform of to ... IA161 Syntactic Formalisms for Parsing Natural Languages 14 / 64 Lecture 7 Description of an object in HPSG sing [ PHON list(phon-string) SYNSEM synsem ] word phrase [ DTRS constituent-struc ] synsem   LOCAL local NON-LOCAL non-local   local      CATEGORY category CONTENT content CONTEXT context      category      HEAD head VAL ... ... ...      IA161 Syntactic Formalisms for Parsing Natural Languages 15 / 64 Lecture 7 Semantic representation: CONTENT (& CONTEXT) feature Semantic interpretation of the sign is given as the value to CONTENT nominal-object: an individual/entity (or a set of them), associated with a referring index, bearing agreement features → INDEX, RESTR Parameterized-state-of-affairs (psoa): a partial situate; an event relation along with role names for identifying the participants of the event→ BACKGR quantifier: some, all, every, a, the, . . . Note: many of these have been reformulated by “Minimal Recursion Semantics (MRS)” which allows underspecification of quantifier scopes. IA161 Syntactic Formalisms for Parsing Natural Languages 16 / 64 Lecture 7 Sub-categorization of content type content ... psoa nom-obj    INDEX index RESTR set(psoa)    laugh‘ [ LAUGHER ref ] give‘       GIVER ref GIVEN ref GIFT ref       drink‘    DRINKER ref DRUNKEN ref    think‘    THINKER ref THOUGHT psoa    . Note: .. ...... Semantic restriction on the index are represented as a value of RESTR. RESTR is an attribute of a nominal object. The value of RESTR is a set of psoa. In turn, RESTR has the attribute of REL whose value can either be referential indices or psoas. IA161 Syntactic Formalisms for Parsing Natural Languages 17 / 64 Lecture 7 Sub-categorization of index type index     PERSON person NUMBER number GENDER gender     referential there it person first second third number singular plural pgender masculine feminine neuter IA161 Syntactic Formalisms for Parsing Natural Languages 18 / 64 Lecture 7 Lexical input of She HEAD VALENCE INDEX RESTR BACKGR noun val ref psoa CASE SUBJ COMPS SPR PER NUM GEND RELN INST nom 3rd sing fem female 1 1 cat ppro context CATEGORY CONTENT CONTEXT LOCALSYNSEM PHON she localsynsemword IA161 Syntactic Formalisms for Parsing Natural Languages 19 / 64 Lecture 7 Lexical input of She sign word phrase PHON SYNSEM list(phon-string) synsem DTRS constituent-struc Each phrase has a DTRS attribute which has a constituent-structure value This DTRS value corresponds to what we view in a tree as daughters (with additional grammatical role information, e.g. adjunct, complement, etc.) By distinguishing different kinds of constituent-structures, we can define different kinds of constructions in a language IA161 Syntactic Formalisms for Parsing Natural Languages 20 / 64 Lecture 7 Structure of phrase head-adj-struc ADJ-DTR ADJ-DTR sign <> head-filler-struc FILL-DTR FILL-DTR sign <> head-mark-struc MARK-DTR MARK-DTR sign <> head-spr-struc SPR-DTR SPR-DTR <> head-subj-struc SUBJ-DTR SUBJ-DTR <> head-comps-struc COMPS-DTR COMP-DTR <> constituent-struc head-struc coord-struc HEAD-DTR CONJ-DTRS CONJUNCTION-DTR word set(sign)sign ... IA161 Syntactic Formalisms for Parsing Natural Languages 21 / 64 Lecture 7 head-subject/complement structure SYNSEM | LOC | CAT DTRS HEAD VAL SUBJ COMPS head-subj-struc PHON SYNSEM SYNSEM | LOC | CAT DTRS HEAD VAL SUBJ COMPS head-comps-struc SYNSEM | LOC | CAT HEAD VAL SUBJ COMPS PHON PHON SYNSEM 3 3 1 1 1 2 VFORM fin3 verb 2 S H H C IA161 Syntactic Formalisms for Parsing Natural Languages 22 / 64 Lecture 7 Questions! (1) How exactly did the last example work? drink has head information specifying that it is a finite verb and subcategories for a subject and an object The head information gets percolated up (the HEAD feature principle) The valence information gets “checked off” as one moves up in the tree (the VALENCE principle) Such principles are treated as linguistic universals in HPSG IA161 Syntactic Formalisms for Parsing Natural Languages 23 / 64 Lecture 7 HEAD-feature principle The value of the HEAD feature of any headed phrase is token-identical with the HEAD value of the head daughter 1 DTRS head-struc SYNSEM | LOC | CAT | HEAD DTRS | HEAD-DTR | SYNSEM | LOC | CAT | HEADphrase 1 IA161 Syntactic Formalisms for Parsing Natural Languages 24 / 64 Lecture 7 VALENCE principle In a headed phrase, for each valence feature F, the F value of the head daughter is the concatenation of the phrase’s F value with the list of F-DTR’s SYNSEM (Pollard and Sag, 1994:348) phrase SS | LOC | CAT | VAL SUBJ COMPS [a] [b] DTRS HEAD-DTR SUBJ-DTR COMP-DTR SS | LOC | CAT | VAL SUBJ [1] [a] COMPS [2],...,[n] [b] SS [1] SS [2] ,..., ss[n] Note: Valence Principle constrains the way in which information is shared between phrases and their head daughters. F can be any one of SUBJ, COMPS, SPR When the F-DTR is empty, the F valence feature of the head daughter will be copied to the mother phrase IA161 Syntactic Formalisms for Parsing Natural Languages 25 / 64 Lecture 7 Questions! (2) Note that agreement is handled neatly, simply by the fact that the SYNSEM values of a word’s daughters are token-identical to the items on the VALENCE lists How exactly do we decide on a syntactic structure? Why the subject is checked off at a higher point in the tree? IA161 Syntactic Formalisms for Parsing Natural Languages 26 / 64 Lecture 7 Immediate Dominance (ID) Principle Every headed phrase must satisfy exactly one of the ID schemata The exact inventory of valid ID schemata is language specific We will introduce a set of ID schemata for English IA161 Syntactic Formalisms for Parsing Natural Languages 27 / 64 Lecture 7 Immediate Dominance (ID) Schemata DTRS head-struc phrase DTRS DTRS DTRS DTRS SS | LOC | CAT | VAL | COMPS head-spr-struc head-marker-struc marker head-adj-struc MARK-DTR | SS | LOC | CAT | HEAD ADJ-DTR | SS | LOC | CAT | HEAD | MOD HEAD-DTR | SS (head-subject) head-comps-struc (head-complement) (head-specifier) (head-marker) (head-adjunct) ... SS | LOC | CAT | VAL | COMPS DTRS head-subj-struc 1 1 IA161 Syntactic Formalisms for Parsing Natural Languages 28 / 64 Lecture 7 head-adjunct structure PHON SS | LOC | CAT DTRS HEAD VAL | SPR head-adj-struc PHON SS | LOC | CAT | HEAD PRD - MOD adj PHON SS LOC | CAT HEAD VAL | SPR noun LOC | CAT | HEAD det 1 2 2 1 3 3 A H IA161 Syntactic Formalisms for Parsing Natural Languages 29 / 64 Lecture 7 Semantic principle The CONTENT value of a headed phrase is token identical to the CONTENT value of the semantic head daughter The semantic head daughter is identified as The ADJ-DTR in a head-adjunct phrase The HEAD-DTR in other headed phrases DTRS phrase head-struc SYNSEM | LOC | CONT DTRS SYNSEM | LOC | CONT DTRS head-adj-struc ADJ-DTR| SYNSEM | LOC | CONT HEAD-DTR | SYNSEM | LOC | CONT (head-adjunct) (non-head-adjunct) 1 1 1 1 head-adj-struc IA161 Syntactic Formalisms for Parsing Natural Languages 30 / 64 Lecture 7 Example 2 Kim likes bagels word PHON SYNSEM Kim LOCAL CAT CONT HEAD SUBJ SPR COMPS ARG-ST INDEX KEY RELS noun ARG 3sg named_rel INST ARG Kim 1 1 2 2 IA161 Syntactic Formalisms for Parsing Natural Languages 31 / 64 Lecture 7 Example 2 Kim likes(1) bagels nowARG2 ARG1 t_overlap_rel ARG2 ARG1 EVENT like_rel RELS KEY INDEX content CONT ARG-ST INDEX content CONT COMPS SPR SUBJ HEAD noun category CAT local LOCAL synsem INDEX content CONT COMPS SPR SUBJ HEAD noun 3sgARG category CAT local LOCAL synsem fin verb FORM HEAD SUBJ SPR COMPS CAT LOCAL SYNSEM category local synsem word PHON likes 3 3 4 5 .6 6 3 5 2 1 2, 4 1 IA161 Syntactic Formalisms for Parsing Natural Languages 32 / 64 Lecture 7 Example 2 Kim likes(2) bagels word PHON SYNSEM likes LOCAL CAT CONT HEAD SUBJ SPR COMPS ARG-ST INDEX RELS 3sgNP NP like_rel EVENT ARG1 ARG2 ARG2 ARG1 t-overlap_rel 3 now 3 4 5 , 3 1 4 52 1 2 FORM verb fin , IA161 Syntactic Formalisms for Parsing Natural Languages 33 / 64 Lecture 7 Example 2 Kim likes bagels word PHON SYNSEM bagels LOCAL CAT CONT HEAD SUBJ SPR COMPS ARG-ST INDEX KEY RELS INST bagel_rel DetP AGR pl noun 1 2 1 2 3 3 IA161 Syntactic Formalisms for Parsing Natural Languages 34 / 64 Lecture 7 Example 2 head-complement schema head-comps-ph PHON SYNSEM HEAD-DTR NON-HEAD-DTRS LOCAL PHON SYNSEM PHON SYNSEM RELS PHON SYNSEM RELS CAT CONT LOCAL CAT CONT HEAD SUBJ SPR COMPS INDEX KEY RELS HEAD SUBJ SPR COMPS INDEX KEY RELS sts Z... N M... D , ... E 2 3 F 1 A B E C 2 3 F M Z... 1 A B C D N... IA161 Syntactic Formalisms for Parsing Natural Languages 35 / 64 Lecture 7 Example 2 head-complement schema headed by likes head-comps-ph PHON SYNSEM HEAD-DTR NON-HEAD-DTRS LOCAL CAT CONT PHON SYNSEM PHON SYNSEM LOCAL | CONT | RELS LOCAL CAT CONT RELS like_rel EVENT ARG1 ARG2 t-overlap_rel ARG1 ARG2 now KEY INDEX COMPS NP SPR SUBJ HEAD NP 3sg likes RELS KEY INDEX COMPS SPR SUBJ HEAD F6 D 22 4 5 ,3E 2 3 6 5 B A 4 verb FORM fin1 A 2 3 E F 1 A B C D IA161 Syntactic Formalisms for Parsing Natural Languages 36 / 64 Lecture 7 Example 2 Kim likes bagels head-comps-ph PHON SYNSEM LOCAL CAT CONT HEAD SUBJ SPR COMPS INDEX KEY RELS FORM NP EVENT ARG1 ARG2 ARG1 ARG2 INST likes, bagels verb fin 3sg like_rel t-overlap_rel now bagel_rel 52 2 4 5 , , 3 2 3 4 IA161 Syntactic Formalisms for Parsing Natural Languages 37 / 64 Lecture 7 Example 2 head-subject schema head-subj-ph PHON SYNSEM HEAD-DTR NON-HEAD-DTRS LOCAL CAT CONT HEAD SUBJ SPR COMPS INDEX KEY RELS PHON SYNSEM LOCAL CAT CONT HEAD SUBJ SPR COMPS INDEX KEY RELS FORM verb fin PHON SYNSEM LOCAL | CONT | RELS F4 B 2 3 E D C 4 1 A 2 3 E F 1 C D B A IA161 Syntactic Formalisms for Parsing Natural Languages 38 / 64 Lecture 7 Example 2 head-subject schema headed by likes bagels NON-HEAD-DTRS PHON SYNSEM LOCAL | CONT | RELS F4 B LOCAL CAT CONT HEAD SUBJ SPR COMPS INDEX KEY RELS FORM NP EVENT ARG1 ARG2 ARG1 ARG2 INST likes, bagels verb fin 3sg like_rel t-overlap_rel now bagel_rel 2 2 5 , , 3 2 3 4 head-subj-ph PHON SYNSEM HEAD-DTR LOCAL CAT CONT HEAD SUBJ SPR COMPS INDEX KEY RELS 2 3 E F 1 C D B A 6 6 E D C 5 1 PHON SYNSEM A IA161 Syntactic Formalisms for Parsing Natural Languages 39 / 64 Lecture 7 Example 2 Kim likes bagels head-subj-ph PHON Kim, likes, bagels SYNSEM LOCAL CAT CONT HEAD SUBJ SPR COMPS INDEX KEY RELS verb finFORM named_rel INST ARG Kim like_rel EVENT ARG1 ARG2 t-overlap_rel ARG1 ARG2 now bagel_rel INST 6 2 2 5 6 5 2 3 , , , IA161 Syntactic Formalisms for Parsing Natural Languages 40 / 64 Lecture 7 Example 2 Tree of Kim likes bagels head-subj-ph verbHEAD SUBJ SPR COMPS word head-comps-ph verbnoun HEAD SUBJ SPR COMPS HEAD SUBJ SPR COMPS HEAD SUBJ SPR COMPS word word verb noun Kim likes bagels 2 2 1 1 1 HEAD SUBJ SPR COMPS IA161 Syntactic Formalisms for Parsing Natural Languages 41 / 64 Lecture 7 Compare HPSG to CFG Each sign or HPSG rule consists of SYNSEM, DTRS, and PHON parts. The SYNSEM part specifies how the syntax and semantics of the phrase (or word) are constrained. It corresponds roughly to the left-hand side of CFG rules but contains much more information. The DTRS part specifies the constituents that make up the phrase (if it is a phrase). (Each of these constituents is a complete sign.) This corresponds to part of the information on the right-hand side of CFG rules, but not to ordering information. The PHON part specifies the ordering of the constituents in DTRS (where this is constrained) and the pronunciation of these (if this is specifiable). This corresponds to the the ordering information on the right-hand side of CFG rules. IA161 Syntactic Formalisms for Parsing Natural Languages 42 / 64 Lecture 7 Simulation of Bottom-up parsing algorithm in HPSG Unify input lexical-signs with lexical-signs in the lexicon. Until no more such unifications are possible Unify instantiated signs with the daughters of instantiated phrasal signs or with phrasal signs in the grammar. . ...... if all instantiated signs but one saturated one (S) are associated with daughters of other instantiated signs and the PHON value of all instantiated signs is completely specified return the complete S structure else fail. IA161 Syntactic Formalisms for Parsing Natural Languages 43 / 64 Lecture 7 Example 2: processing of unification Kim walks The words in the sentence specify only their pronunciations and their positions. 1 [PHON (( 0 1 kim))] 2 [PHON (( 1 2 walks))] . STEP 1: Unifying 1 with the lexical entry for Kim gives .. ...... 3 [PHON ((0 1 kim)) SYNSEM [CAT [HEAD noun SUBCAT ()] CONTENT [INDEX 1 [PER 3rd NUM sing]] CONTEXT [BACKGR {[RELN naming BEARER 1 NAME Kim]}]]] We now know something about the meaning of Kim (it refers to somebody named Kim) and something about its syntactic properties (it is third person singular). IA161 Syntactic Formalisms for Parsing Natural Languages 44 / 64 Lecture 7 Example 2: processing of unification 1 [PHON (( 0 1 kim))] 2 [PHON (( 1 2 walks))] . STEP 2: Unifying 2 with the lexical entry for walks gives .. ...... 4 [PHON ((1 2 walks)) SYNSEM [CAT [HEAD [VFORM fin] SUBCAT ([CAT [HEAD noun SUBCAT ()] CONTENT [INDEX 1 [PER 3rd NUM sing]]])] CONTENT [RELN walk WALKER 1]]] We know that walks refers to walking and that it requires a subject noun phrase which refers to the walker but doesn’t require any object. IA161 Syntactic Formalisms for Parsing Natural Languages 45 / 64 Lecture 7 Example 2: processing of unification . HEAD-DTR rule .. ...... [SYNSEM [CAT [HEAD 1 SUBCAT (2)] CONTENT 4] DTRS [HEAD-DTR [SYNSEM [CAT [HEAD 1 SUBCAT (2)] CONTENT 4] PHON 3] SUBJ-DTRS ()] PHON 3] . STEP 3: Unifying 4 with the HEAD-DTR of this rule gives .. ...... 5 [SYNSEM [CAT [HEAD [VFORM fin] SUBCAT 2([CAT [HEAD noun SUBCAT ()] CONTENT [INDEX 1 [PER 3rd NUM sing]]])] CONTENT 4[RELN walk WALKER 1]] DTRS [HEAD-DTR [SYNSEM [CAT [HEAD [VFORM fin] SUBCAT (2)]] CONTENT [4] PHON 3((1 2 walks))] SUBJ-DTRS ()] PHON 3((1 2 walks))] Now we have a VP with the transitive verb walks as its head (and only constituent). IA161 Syntactic Formalisms for Parsing Natural Languages 46 / 64 Lecture 7 Example 2: processing of unification . HEAD-DTR rule .. ...... 6 [SYNSEM [CAT [HEAD 1 SUBCAT ()] CONTENT 4] DTRS [HEAD-DTR [SYNSEM [CAT [HEAD 1 SUBCAT (2)] CONTENT 4] PHON 3] SUBJ-DTRS ([PHON 5 SYNSEM 2])] PHON (5 < 3)] . STEP 4: Unifying 5 with the HEAD-DTR of this rule gives .. ...... 7 [SYNSEM [CAT [HEAD 1[VFORM fin SUBCAT ()]] CONTENT 4[RELN walk WALKER ]] DTRS [HEAD-DTR [SYNSEM [CAT [HEAD 1[VFORM fin] SUBCAT 2([CAT [HEAD noun SUBCAT ()] CONTENT [INDEX [PER 3rd NUM sing]]])] CONTENT [RELN walk WALKER 4]] PHON 3((1 2 walks))] SUBJ-DTRS ([PHON 5 SYNSEM 2[CAT [HEAD noun SUBCAT ()] CONTENT [INDEX ]]])] PHON ( 5 < 3((1 2 walks)))] IA161 Syntactic Formalisms for Parsing Natural Languages 47 / 64 Lecture 7 Example 2: processing of unification . STEP 5: Unifying 3 with the SUBJ-DTR of 7 gives .. ...... 8 [SYNSEM [CAT [HEAD [VFORM fin SUBCAT ()]] CONTENT [RELN walk WALKER [PER 3rd NUM sing]]] DTRS [HEAD-DTR [SYNSEM [CAT [HEAD [VFORM fin] SUBCAT ([CAT [HEAD noun SUBCAT ()] CONTENT [INDEX [PER 3rd NUM sing]]]) CONTENT [RELN walk WALKER [PER 3rd NUM sing]]] PHON ((1 2 walks))] SUBJ-DTRS ([PHON ((0 1 kim)) SYNSEM [CAT [HEAD noun SUBCAT ()] CONTENT [INDEX [PER 3rd NUM sing]]])] PHON ((0 1 kim) (1 2 walks))] Now the subject of the sentence is pronounceable, and we’re done. IA161 Syntactic Formalisms for Parsing Natural Languages 48 / 64 Lecture 7 Phenomena covered by HPSG parsers Case assignment Word order : scrambling Long distance dependency Coordination Scope of adverbs and negation Topic drop Agreement Relative clause … IA161 Syntactic Formalisms for Parsing Natural Languages 49 / 64 Lecture 7 Example 3: unbounded dependency construction An unbounded dependency construction involves constituents with different functions involves constituents of different categories is in principle unbounded Two kind of unbounded dependency constructions (UDCs) Strong UDCs Weak UDCs IA161 Syntactic Formalisms for Parsing Natural Languages 50 / 64 Lecture 7 Strong UDCs An overt constituent occurs in a non-argument position: Topicalization: Kimi, Sandy loves_ i. Wh-questions: I wonder [whoi Sandy loves_ i]. Wh-relative clauses: This is the politician [whoi Sandy loves_ i]. It -clefts: It is Kim i [whoi Sandy loves_ i]. Pseudoclefts: [Whati Sandy loves_ i ] is Kimi. IA161 Syntactic Formalisms for Parsing Natural Languages 51 / 64 Lecture 7 Weak UDCs No overt constituent in a non-argument position: Purpose infinitive (for -to clauses): I bought iti for Sandy to eat_ i . Tough movement: Sandyi is hard to love_ i . Relative clause without overt relative pronoun: This is [the politician]i [Sandy loves_ i ]. It-clefts without overt relative pronoun: It is Kimi [Sandy loves_ i ]. IA161 Syntactic Formalisms for Parsing Natural Languages 52 / 64 Lecture 7 Using the feature SLASH To account for UDCs, we will use the feature SLASH (so-named because it comes from notation like S/NP to mean an S missing an NP) This is a non-local feature which originates with a trace, gets passed up the tree, and is finally bound by a filler IA161 Syntactic Formalisms for Parsing Natural Languages 53 / 64 Lecture 7 The bottom of a UDC: Traces word PHON SYNSEM LOCAL NONLOC INHERITED | SLASH TO-BIND | SLASH 1 1 phonologically null, but structure-shares local and slash values IA161 Syntactic Formalisms for Parsing Natural Languages 54 / 64 Lecture 7 Traces Because the local value of a trace is structure-shared with the slash value, constraints on the trace will be constraints on the filler. For example, hates specifies that its object be accusative, and this case information is local So, the trace has [synsem|local|cat|head|case acc] as part of its entry, and thus the filler will also have to be accusative *Hei/Himi, John likes_ i IA161 Syntactic Formalisms for Parsing Natural Languages 55 / 64 Lecture 7 The middle of a UDC: The Nonlocal Feature Principle (NFP) For each NON-LOCAL feature, the inherited value on the mother is the union of the inherited values on the daughter minus the to-bind value on the head daughter. In other words, the slash information (which is part of inherited) percolates “up” the tree This allows the all the local information of a trace to “move up” to the filler IA161 Syntactic Formalisms for Parsing Natural Languages 56 / 64 Lecture 7 The middle of a UDC: The Nonlocal Feature Principle (NFP) The top of a UDC: filler-head structures Example for a structure licensed by the filler-head schema NLOC | INHERITED | SLASH LOCAL NLOC INHERITED | SLASH TO-BIND | SLASH 1 1..., ,...1 F H IA161 Syntactic Formalisms for Parsing Natural Languages 57 / 64 Lecture 7 The middle of a UDC: The Nonlocal Feature Principle (NFP) The analysis of the UDC example Johni we know She likes_i S NLOC | INHERITED | SLASH NLOC INHERITED | SLASH TO-BIND | SLASH S VP NLOC INHERITED | SLASH TO-BIND | SLASH S LOC | CAT | SUBCAT NLOC INHERITED | SLASH TO-BIND | SLASH VP LOC | CAT | SUBCAT NLOC INHERITED | SLASH TO-BIND | SLASH V LOC | CAT | SUBCAT NONLOC | TO-BIND | SLASH LOC NLOC | INHER | SLASH NP NP V NP John we know she likes -i 3 1 1 2 2 3 3 3 3 3 3 NP , HF LOCAL 3 i HS CH HS H C 1 IA161 Syntactic Formalisms for Parsing Natural Languages 58 / 64 Lecture 7 Example 4 John reads a new book PHON SYNSEM | LOC CAT CONT READER READEE HEAD VAL VFORM AUX INV SUBJ COMPS SPR NP NP reads fin bool bool nom,-PRD acc,-PRD word read verb 3rd,sg 1 2 1 2 IA161 Syntactic Formalisms for Parsing Natural Languages 59 / 64 Lecture 7 Example 4 John reads a new book PHON SYNSEM | LOCAL CAT | HEAD CONT MOD LOCAL PRD CAT CONT HEAD VAL | SPR INDEX RESTR INDEX RESTR RELN ARG new noun new cat nom-obj localsynsem adj nom-obj local word - 1 1 1 2 2 IA161 Syntactic Formalisms for Parsing Natural Languages 60 / 64 Lecture 7 Example 4 John reads a new book Note: apply head-adjunct schema PHON SS | LOC CAT CONT HEAD VAL | SPR PHON SS | LOC CAT | HEAD | MOD CONT INDEX RESTR RELN ARG PHON SS LOC CAT CONT HEAD VAL | SPR INDEX RESTR PER NUM GEN RELN INST book neut sg 3rd book new new book nom-obj nom-obj new 1 1 2 2 3 4 4 4 4 3 5 5 A H IA161 Syntactic Formalisms for Parsing Natural Languages 61 / 64 Lecture 7 Example 4 John reads a new book PHON SS | LOC CAT CONT HEAD VAL | SPR PHON SS LOC | CAT | HEAD SPEC PHON SS LOC CAT CONT HEAD VAL | SPR a new book a new book det 1 1 2 2 6 67 7 SPR H IA161 Syntactic Formalisms for Parsing Natural Languages 62 / 64 Lecture 7 Example 4 John reads a new book PHON SS | LOC CAT CONT HEAD VAL SUBJ COMPS NP nom,-PRD PHON SS | LOC CAT CONT HEAD VAL VFORM SUBJ COMPS NP acc,-PRD READER READEE PHON SS LOC CAT CONT | INDEX HEAD VAL CASE PRD SUBJ COMPS SPR reads a new book reads a new book 3rd,sg fin acc noun read verb 7 7 4 4 4 8 8 9 9 10 10 11 11 H C IA161 Syntactic Formalisms for Parsing Natural Languages 63 / 64 Lecture 7 Example 4 John reads a new book - completed analysis PHON SS | LOC CAT CONT HEAD VAL SUBJ COMPS SPR PHON SS LOC CAT CONT HEAD VAL INDEX RESTR CASE PRD - SUBJ COMPS SPR PER NUM GEND NAME INST PHON SS | LOC CAT CONT HEAD VAL SUBJ COMPS READER READEE John reads a new book John reads a new book 4 7 7 8 8 9 9 11 11 11 nom 3rd sg masc John naming nom-obj read noun SUBJ H IA161 Syntactic Formalisms for Parsing Natural Languages 64 / 64