Inteligencia Artificial Interpretacin semntica Primavera 2009 profesor Luigi

Inteligencia Artificial Interpretación semántica Primavera 2009 profesor: Luigi Ceccaroni

CFG recognition using difference lists • An efficient implementation of CFGs can be obtained by making use of difference lists: a sophisticated Prolog technique. • The key idea underlying difference lists is to represent the information about grammatical categories not as a single list, but as the difference between two lists. • For example, instead of representing a woman shoots a man as [a, woman, shoots, a, man] we might represent it as the pair of lists [a, woman, shoots, a, man] [ ].

CFG recognition using difference lists • Think of the first list as what needs to be consumed (or if you prefer: the input list), and the second list as what we should leave behind (or: the output list). • Viewed from this (rather procedural) perspective the difference list [a, woman, shoots, a, man] [ ]. • represents the sentence a woman shoots a man because it says: If I consume all the symbols on the left, and leave behind the symbols on the right, I have the sentence I am interested in.

CFG recognition using difference lists • The sentence we are interested in is the difference between the contents of the two lists. • Difference representations are not unique. • In fact, we could represent a woman shoots a man in infinitely many ways. • For example, we could also represent it as [a, woman, shoots, a, man, ploggle, woggle] [ploggle, woggle]. • Again the point is: if we consume all the symbols on the left, and leave behind the symbols on the right, we have the sentence we are interested in.

CFG recognition using difference lists • If we bear the idea of consuming something, and leaving something behind in mind, we obtain the following recognizer (Prolog notation): s(X, Z) : - np(X, Y), vp(Y, Z). np(X, Z) : - det(X, Y), n(Y, Z). vp(X, Z) : - v(X, Y), np(Y, Z). vp(X, Z) : - v(X, Z). det([the|W], W). det([a|W], W). n([woman|W], W). n([man|W], W). v([shoots|W], W). The s rule says: I know that the pair of lists X and Z represents a sentence if (1) I can consume X and leave behind a Y, and the pair X and Y represents a noun phrase, and (2) I can then go on to consume Y leaving Z behind, and the pair Y Z represents a verb phrase.

CFG recognition using difference lists • The idea underlying the way we handle the words is similar. • The code n([man|W], W). • means we are handling man as the difference between [man|W] and W. • Intuitively, the difference between what I consume and what I leave behind is precisely the word man.

CFG recognition using difference lists • How do we use such grammars? Here's how to recognize sentences: s([a, woman, shoots, a, man], []). yes • This asks whether we can get an s by consuming the symbols in [a, woman, shoots, a, man], leaving nothing behind.

CFG recognition using difference lists • Similarly, to generate all the sentences in the grammar, we ask s(X, []). • This asks: what values can you give to X, such that we get an s by consuming the symbols in X, leaving nothing behind? • The queries for other grammatical categories also work the same way. • For example, to find out if a woman is a noun phrase we ask: np([a, woman], []).

CFG recognition using difference lists • And we generate all the noun phrases in the grammar as follows: np(X, []). • It has to be admitted that this recognizer is not as easy to understand, at least at first, and it's a pain having to keep track of all those difference list variables. • This is where DCGs come in.

Formalismos de unificación • La unificación se usa como mecanismo básico de composición entre constituyentes en gramáticas lógicas. • Història: – Q-Systems (Colmerauer, 1972) – Prolog (Colmerauer, 1973) – Gramàtiques de Metamorfosi (Colmerauer, 1975) – Gramàtiques de Clàusules Definides (DCGs) (Pereira, Warren, 1980)

Análisis gramatical con unificación • Gramática (1) oració (X, Y) (2) gnom(X, Y) (3) gver(X, Y) • Lexicón (4) art(X, Y) (5) nom(X, Y) (6) ver(X, Y) : - gnom(X, Z), gver(Z, Y) : - art(X, Z), nom(Z, Y) : - ver(X, Y) : - el(X, Y) : - gos(X, Y) : - borda(X, Y)

Análisis gramatical con unificación 1 el 2 gos 3 borda (7) el(1, 2) (8) gos(2, 3) (9) borda(3, 4) Frase a analizar: oració(1, 4) 4

Análisis gramatical con unificación 1 el 2 gos 3 borda 4 oració(1, 4) (R 1) (X 1, Y 4) por unificación gnom (1, Z), gver(Z, 4) (R 2) aplicada a gnom(1, Z) art(1, U), nom(U, Z), gver(Z, 4)

Análisis gramatical con unificación 1 el 2 gos 3 borda 4 art(1, U), nom(U, Z), gver(Z, 4) (R 4) aplicada a art(1, U) el(1, U), nom(U, Z), gver(Z, 4) (R 7) (U 2) el(1, 2), nom(2, Z), gver(Z, 4)

Análisis gramatical con unificación 1 el 2 gos nom(2, Z), gver(Z, 4) (R 5) gos(2, Z), gver(Z, 4) (R 8) (Z 3) gos(2, 3), gver(3, 4) 3 borda 4

Análisis gramatical con unificación 1 gver(3, 4) (R 3) ver(3, 4) (R 6) borda(3, 4) (R 9): Fin el 2 gos 3 borda 4

Análisis semántico • Consiste en construir una representación de las frases en algún sistema formal. • En general es un problema intratable. • Se simplifica suponiendo que la semántica de una frase se pueda construir a partir de la semántica de sus partes: semántica compositiva. • Algunas características del lenguaje se tienen que tratar a parte: referencias, omisiones, contexto…

Estrategias de análisis • Dos maneras de plantear la interpretación semántica: – Secuencial (sintáctica → semántica) – Paralela (sintáctica + semántica)

Interpretación secuencial • Problemas de ambigüedad – Es posible que haya más de una interpretación sintáctica. – Hay que considerarlas todas, para poder comprobar sucesivamente cuáles son las semánticamente posibles. • Principal ventaja – El análisis semántico parte de un análisis sintáctico correcto.

Interpretación paralela • Principal problema – No se sabe si la interpretación sintáctica es correcta hasta el final. • Principal ventaja – Poder descartar interpretaciones sintácticas correctas (o parcialmente correctas) que no tengan interpretación semántica asociada. • Las reglas sintácticas incluyen la información semántica asociada. – Se obtiene como resultado un árbol de análisis y una o varias interpretaciones.

Sistema de representación • El sistema de representación tiene que permitir: – Manejar cuantificación, predicación, negación, modalidad (creencias) – Resolver la ambigüedad tanto léxica (polisemia) como sintáctica – Manejar inferencias (herencia, razonamiento por omisión) • Importante a la hora de resolver problemas que involucren el contexto o el conocimiento del dominio.

Sistema de representación • Los sistemas que no se basan en la sintaxis para la interpretación suelen utilizar sistemas tipo ontologías. • Por lo general, se utiliza una variedad de la lógica de primer orden adecuada al dominio de aplicación. • El elemento básico de representación es el lexema: raíz de un grupo de palabras que son diferentes formas de “la misma palabra” (ej. : ir, ido, yendo).

Sistema de representación • Ejemplos – Los nombres propios corresponden a constantes. – Los verbos intransitivos a predicados unarios: Juan ríe(juan) – Los verbos transitivos a predicados de aridad superior: Juan lee el Quijote lee(juan, quijote) – Los nombres genéricos a predicados sobre variables: El hombre(X) – Los adjetivos a predicados unarios: La casa grande(X) ∧ casa(X)

Sistema de representación La representación consiste normalmente en un árbol de análisis y una función de composición que construye la interpretación de los elementos asociados. S sem. S = f. S (sem. NP , sem. VP) VP NP sem. NP = f. NP (sem. ART , sem. N) sem. ART = f. ART (sem el) NAME Adrià NP V menja ART N el bacallà

La representación del significado: ejemplo entrada: Qui dirigeix el PSOE? forma lògica: (pregunta (referent (X)) (X instancia (X, persona) (el 1 (Y instancia(Y, partit_polític) nom(Y, "PSOE")) (Z instancia(Z, dirigir) present(Z) valor_prop(Z, agent, X) valor_prop(Z, pacient, Y)))))

La representación del significado • En la anterior forma lógica aparecen 4 tipos diferentes de información: – Estructura lógica – Contenido conceptual (semántico) – Indicación de los actos del discurso – Anotaciones pragmáticas • El formalismo de representación debería proporcionar una capacidad expresiva suficiente para garantizar la descripción de estos 4 tipos de información.

Posibles representaciones del significado • Lógica de predicados de primer orden • Otros formalismos lógicos • Ontologías 27

Representaciones basadas en lógica • Un vocabulario de predicados de los que hay que indicar la aridad (el número de argumentos y su tipo) • Un vocabulario de constantes y variables. • Un conjunto de conectores lógicos. • Un vocabulario de funciones de las que hay que indicar la aridad • Un conjunto de cuantificadores que actúan sobre los predicados que pueden ser 28 cuantificados.

Semántica y forma lógica • Semantics = theories of meaning • Introduction of a level of contextindependent meaning called the logical form. – It can be produced directly from the syntactic structure of a sentence. – It does not contain the results of any analysis that requires interpretation of the sentence in context. 29

Semántica y forma lógica • Precisely defining the notions of semantics and meaning is surprisingly difficult. – The terms are used for several different purposes in natural and technical usage. – There is a use of the verb mean that has nothing to do with language: This fire means someone camped here last night. 30

Semántica y forma lógica This fire means someone camped here last night. • In this example, the fire is evidence for (or implies) the conclusion. • The meaning we are interested in is not this one. • It is closer to the usage when defining a word. 31

Semántica y forma lógica “Amble” means to walk slowly. • This defines the meaning of a word in terms of other words. • We are interested in specifying meaning without having to refer back to natural language itself. • But even if we can do this, defining a notion of sentence meaning is difficult. 32

Semántica y forma lógica A guard at the airport: “Do you know what gate you are going to? ” • If we interpret this as asking whether we know where we are going, we answer yes. • But the response can be based on a misunderstanding, if the guard then asks: “Which gate is it? ” 33

Semántica y forma lógica “Do you know what gate you are going to? ” • This sentence then appears to mean different things in different contexts. • Can we define a notion of sentence meaning that is independent of context? • Is there a level at which this sentence has a single meaning, but may be used for different purposes? 34

Semántica y forma lógica • Separating sentence meaning from sentence usage is complex, but there are many advantages, primarily modularity. • If sentences have no context-independent meaning, then we may not be able to separate: – the study of language – the study of general human reasoning and context 35

Semántica y forma lógica • We use the term meaning in a contextindependent sense. • We use the term usage for the contextdependent aspects. • The representation of context-independent meaning is called the logical form. 36

Semántica y forma lógica • The process of mapping a sentence to its logical form is called semantic interpretation. • The process of mapping the logical form to the final knowledge representation language is called contextual interpretation. 37

Semántica y forma lógica S Syntactic analysis VP NP ART N la pelota V es ADJP roja semantic interpretation Logical form (ROJA 1 <LA b 1 PELOTA>) contextual interpretation Final representation Roja(BO 73) 38

Interpretación semántica: función de composición • Interpretación a través de evaluaciones lambda: (lambda (x) (. . . )) = = ( (x) (. . . )) • Gramática: Oracio GN FV (2 1) GN np (1) FV vi (1) Fv vt GN (1 2) • Lexicón: Pere np, pere Maria np, maria riu vi, ( (x) (riu(x)) estima vt , (( (x) ( (y), estima(y, x)))) 39

Función de composición: ejemplo 40

Función de composición: ejemplo 41

Gramáticas de cláusulas definidas (DCGs) • Las DCGs permiten escribir gramáticas en forma de programas como, por ejemplo, Prolog. • Prolog es un lenguaje de reglas que usa razonamiento hacia atrás como método de resolución. • Se define una sintaxis especial que permite diferenciar los elementos gramaticales de los procedimientos que se usan para tener en cuenta el contexto. 42

Gramáticas de cláusulas definidas (DCGs): sintaxis • Las reglas usan variables para pasarse información, hacer las comprobaciones que exige la gramática o construir un resultado. • Se puede introducir código Prolog, entre llaves: aaa(W) --> [W], bbb(W), {number(W)} 43

Gramáticas de cláusulas definidas (DCGs): sintaxis • Para ejecutar una DCG se llama al símbolo principal de la gramática con dos parámetros: una lista con las palabras de la frase y una lista vacía: frase([el, gat, menja, bacalla], []) 44

Gramáticas de cláusulas definidas (DCGs): ejemplo analisis(X, Y): - asercion(X, Y). asercion --> sn, verb, compl --> []. compl --> prep, sn. compl--> sn. sn--> npr. sn--> det, n. verb--> [W], {verbo(W)}. npr--> [W], {npropio(W)}. n--> [W], {nombre(W)}. det--> [W], {determ(W)}. prep--> [W], {prepo(W)}. npropio(clara). npropio(maria). npropio(barcelona). nombre(hombre). nombre(profesor). nombre(libro). determ(un). determ(el). verbo(esta). verbo(rie). verbo(piensa). verbo(habla). verbo(lee). prepo(en). prepo(con). prepo(de). 45

What are definite clause grammars? • Definite Clause Grammars (DCGs) are convenient ways to represent grammatical relationships for various parsing applications. • They can be used for natural language work, for creating formal command programming languages. • Quite simply, they are a nice notation for writing grammars that hides the underlying difference list variables.

DCGs • A little grammar written as a DCG: s --> np, vp. np --> det, n. vp --> v, np. vp --> v. det --> [the]. det --> [a]. n --> [woman]. n --> [man]. v --> [shoots]. • How do we use this DCG? In fact, we use it in exactly the same way as we used the difference list recognizer.

DCGs • For example, to find out whether a woman shoots a man is a sentence, we pose the query: s([a, woman, shoots, a, man], []). • That is, just as in the difference list recognizer, we ask whether we can get an s by consuming the symbols in [a, woman, shoots, a, man], leaving nothing behind.

DCGs • Similarly, to generate all the sentences in the grammar, we pose the query: s(X, []). • This asks what values we can give to X, such that we get an s by consuming the symbols in X, leaving nothing behind. • Moreover, the queries for other grammatical categories also work the same way. For example, to find out if a woman is a noun phrase we pose the query: np([a, woman], []).

DCGs • We generate all the noun phrases in the grammar as follows: np(X, []). • Quite simply, this DCG is a difference list recognizer! • That is, DCG notation is essentially syntactic sugar: user friendly notation that lets us write grammars in a natural way.

DCGs • The Prolog language can translate this notation into the kinds of difference lists discussed before. • So we have the best of both worlds: – a nice simple notation for working with – the efficiency of difference lists

DCGs • To see what Prolog translates DCG rules into: – let Prolog consult the rules of this DCG, then if you pose the query: listing(s) – you will get the response: s(A, B) : np(A, C), vp(C, B). • This is what Prolog has translated s --> np, vp into. • Note that this is exactly the difference list rule we used in the recognizer.

DCGs • Similarly, if you pose the query: listing(np) • you will get: np(A, B) : det(A, C), n(C, B). • This is what Prolog has translated np --> det, n into. • Again (apart from the choice of variables) this is the difference list rule we used in the recognizer.

DCGs • To get a complete listing of the translations of all the rules, simply type: listing.

Separating rules and lexicon • By separating rules and lexicon we mean that we want to eliminate all mentioning of individual words in the DCGs and instead record all the information about individual words separately in a lexicon. • To see what is meant by this, let's return to the basic grammar, namely: np - - > det, n. vp - - > v, np. vp - - > v. det - - > [the]. det - - > [a]. n - - > [woman]. n - - > [man]. v - - > [shoots].

Separating rules and lexicon • We are going to separate the rules form the lexicon. • That is, we are going to write a DCG that generates exactly the same language, but in which no rule mentions any individual word. • All the information about individual words will be recorded separately.

Separating rules and lexicon • Here is an example of a (very simple) lexicon. • Lexical entries are encoded by using a predicate lex/2 whose first argument is a word, and whose second argument is a syntactic category: lex(the, det). lex(a, det). lex(woman, n). lex(shoots, v).

Separating rules and lexicon • A simple grammar that could go with this lexicon will be very similar to the basic DCG. • In fact, both grammars generate exactly the same language. • The only rules that change are those that mention specific words, i. e. the det, n, and v rules. det --> [Word], {lex(Word, det)}. n --> [Word], {lex(Word, n)}. v --> [Word], {lex(Word, v)}.

Separating rules and lexicon Grammar: np - - > det, n. vp - - > v, np. vp - - > v. det --> [Word], {lex(Word, det)}. n --> [Word], {lex(Word, n)}. v --> [Word], {lex(Word, v)}.

Separating rules and lexicon • Consider the new det rule: det --> [Word], {lex(Word, det)}. • This rule says “a det can consist of a list containing a single element Word” (note that Word is a variable). • The extra test adds the crucial condition: “as long as Word matches with something that is listed in the lexicon as a determiner”.

Separating rules and lexicon • With our present lexicon, this means that Word must be matched either with the word “a” or “the”: lex(the, det). lex(a, det). • So this single rule replaces the two previous DCG rules for det.

Separating rules and lexicon • This explains the how of separating rules from lexicon, but it doesn't explain the why. • Is it really so important? • Is this new way of writing DCGs really that much better?

Separating rules and lexicon • The answer is yes! for a theoretical reason: – Arguably rules should not mention specific lexical items. – The purpose of rules is to list general syntactic facts, such as the fact that a sentence can be made up of a noun phrase followed by a verb phrase. – The rules for s, np, and vp describe such general syntactic facts, but the old rules for det, n, and v don't. – Instead, the old rules simply list particular facts: that a is a determiner, that the is a determiner, and so on. – From a theoretical perspective it is much neater to have a single rule that says “anything is a determiner (or a noun, or a verb, . . . ) if it is listed as such in the lexicon”.

Separating rules and lexicon • Now, our little lexicon, with its simple lex entries, is a toy. • But a real lexicon is (most emphatically!) not. • A real lexicon is likely to be very large (it may contain hundreds of thousands, or even millions, of words) and moreover, the information associated with each word is likely to be very rich.

Separating rules and lexicon • Our lex entries give only the syntactical category of each word. • A real lexicon will give much more, such as information about its phonological, morphological, semantic, and pragmatic properties. • Because real lexicons are big and complex, from a software engineering perspective it is best to write simple grammars that have a well-defined way of pulling out the information they need from vast lexicons.

Separating rules and lexicon • That is, grammars should be thought of as separate entities which can access the information contained in lexicons. • We can then use specialized mechanisms for efficiently storing the lexicon and retrieving data from it. • The new rules really do just list general syntactic facts, and the extra tests act as an interface to our (admittedly simple) lexicon that lets the rules find exactly the information they need.

Grammar 1: a trivial grammar for a fragment of language s np, vp. % A sentence (s) is a noun phrase (np) plus a verb phrase (vp) np det, n. % A noun phrase is a determiner plus a noun np n. %. . . or just a noun. vp v, np. % A verb phrase is a verb and its direct object, which is an np vp v. %. . . or just the verb (for intransitives). det [Word], {lex(Word, det)}. n [Word], {lex(Word, n)}. v [Word], {lex(Word, v)}.

Grammar 1: a trivial grammar for a fragment of language lex(the, det). % ‘the’ is a determiner lex(mary, n). % ‘mary’ is a noun. lex(john, n). lex(woman, n). lex(apple, n). lex(man, n). lex(loves, v). % ‘loves’ is a verb. lex(eats, v). lex(sings, v).

Sentences for Grammar 1 • • mary loves john the woman eats the apple the man sings mary eats

Grammar 2: restrictions in argument selection s np, vp. compl([ ]) [ ]. compl([arg(X)]) p(X), np. compl([ ]) np. np name. np det, n. vp v(X), compl(X). % A vp is a verb plus a verbal complement (compl)

Grammar 2: restrictions in argument selection v(A) [Word], {lex(Word, v, A)}. name [Word], {lex(Word, name)}. n [Word], {lex(Word, n)}. det [Word], {lex(Word, det)}. p(Word) [Word], {lex(Word, p)}.

Grammar 2: restrictions in argument selection lex(piensa, v, [arg(en)]). lex(está, v, [arg(en)]). lex(ríe, v, [ ]). lex(habla, v, [arg(con)]). lex(lee, v, [ ]). lex(el, det). % ‘el’ is a determiner

Grammar 2: restrictions in argument selection lex(un, det). lex(mary, name). lex(john, name). lex(profesor, n). lex(en, p)

Sentences for Grammar 2 • • mary piensa en john habla con mary john ríe un profesor habla con mary

Extension of Grammar 2 John habla de Clara con Mary • Needed modifications: compl([arg(X) | Y]) p(X), np, compl(Y). lex(habla, v, [arg(de), arg(con)]). lex(clara, name).

Grammar 3: logical representation of sentences s(F) np(S), v(S, X, F), compl(X). compl([ ]) [ ]. compl([arg(X, O) | Y]) p(X), np (O), compl(Y). compl([arg(null, O) | Y]) np(O), compl(Y). np(S) name(S). np(S) det, n(S).

Grammar 3: logical representation of sentences v(S, A, F) [Word], {lex(Word, v, S, A, F)}. name(Word) [Word], {lex(Word, name)}. n(Word) [Word], {lex(Word, n)}. det [Word], {lex(Word, det)}. p(Word) [Word], {lex(Word, p)}.

Grammar 3: logical representation of sentences lex(clara, name). lex(maria, name). lex(juan, name). lex(barcelona, name). lex(libro, n). lex(hombre, n). lex(profesor, n).

Grammar 3: logical representation of sentences lex(el, det). lex(un, det). lex(en, p). lex(con, p). lex(de, p).

Grammar 3: logical representation of sentences lex(ríe, v, S, [ ], reir(S)). lex(piensa, v, S, [arg(en, O)], pensar_en(S, O)). lex(habla, v, S, [arg(de, O), arg(con, O 1)], comunica(S, O, O 1)). lex(habla, v, S, [arg(con, O), arg(de, O 1)], comunica(S, O 1, O)). lex(está, v, S, [arg(en, O)], locativo(S, O)). lex(lee, v, S, [arg(null, O)], leer(S, O)).

Sentences for Grammar 3 • unary predicate (ríe, v, S, [ ], reir(S)). • binary predicate (piensa, v, S, [arg(en, O)], pensar_en(S, O)). • ternary predicate (habla, v, S, [arg(con, O), arg(de, O 1)], comunica(S, O 1, O)). • Example: Juan piensa en Maria = pensar_en(juan, maria).

Prolog input and output analysis(F, X, [ ]): - s(F, X, [ ]). | ? - analysis(F, [juan, está, en, barcelona], [ ]). F = locativo(juan, barcelona) ? yes | ? - analysis(F, [juan, piensa, en, maria], [ ]). F = pensar_en(juan, maria) ? yes

Prolog input and output | ? - analysis(F, [el, libro, está, en, barcelona], [ ]). F = locativo(libro, barcelona) ? yes | ? - analysis(F, [juan, lee, un, libro], [ ]). F = leer(juan, libro) ? yes

Prolog input and output | ? - analysis(F, [el, hombre, habla, de, juan, con, maria], [ ]). F = comunica(hombre, juan, maria) ? yes | ? - analysis(F, [el, hombre, ríe], [ ]). F = reir(hombre) ? yes

Prolog input and output | ? - analysis(F, [el, profesor, piensa, en, un, libro], [ ]). F = pensar_en(profesor, libro) ? yes

Grammar 4: quantification • It exists X and X is a libro: el libro = e(X, libro(X)). • All X such that X is a libro: todo libro = a(X, libro(X)).

Grammar 4: quantification el libro cae = e(X, and(libro(X), cae(X))) Juan piensa en el libro = e(X, and(libro(X), piensa(juan, X))) todo hombre piensa en el libro = a(X, implies(hombre(X), e(Y, and(libro(Y), piensa(X, Y)))

Grammar 4: quantification lex(el, det, K, S 1, S 2, e(K, and(S 1, S 2))). lex(un, det, K, S 1, S 2, e(K, and(S 1, S 2))). lex(los, det, K, S 1, S 2, a(K, implies(S 1, S 2))). lex(todo, det, K, S 1, S 2, a(K, implies(S 1, S 2))). np(K, S 2, F) det(K, S 1, S 2, F), n(K, S 1). np(K, F, F) name(K).

Grammar 4: quantification compl([ ], S, S) [ ]. compl([arg(X, K) | Y], S 1, S) p(X), np(K, S 2, S), compl(Y, S 1, S 2). s(S) np(K, S 2, S), v(K, X, S 1), compl(X, S 1, S 2). n(K, F) [Word], {lex(Word, n, K, F)}. lex(libro, n, K, libro(K)).

Prolog input and output | ? - analysis(F, [el, hombre, ríe], [ ]). F = e(_A, and(hombre(_A), reir(_A))) ? yes | ? - analysis(F, [el, profesor, piensa, en, un, libro], [ ]). F = e(_B, and(profesor(_B), e(_A, and(libro(_A), pensar_en(_B, _A))))) ? yes

Prolog input and output | ? - analysis(F, [el, hombre, habla, de, juan, con, maria], [ ]). F = e(_A, and(hombre(_A), comunica(_A, juan, maria))) ? yes | ? - analysis(F, [todo, hombre, piensa, en, un, libro], [ ]). F = a(_B, implies(hombre(_B), e(_A, and(libro(_A), pensar_en(_B, _A))))) ? yes

Prolog input and output | ? - analysis(F, [todo, libro, esta, en, barcelona], [ ]). F = a(_A, implies(libro(_A), locativo(_A, barcelona))) ? yes

Extension of Grammar 4 El hombre bueno: e(X, and(hombre(X), bueno(X)))

Grammar 5: semantic constraints • Argument selection: – *Juan lee el hombre – El gato come pescado – El perro corre por el camino • Semantic categories: – human: Juan, María, Clara, hombre, profesor – animate: perro, gato – inanimate: libro, pescado, bocadillo, silla – locative: Barcelona, camino

Nouns’ semantic categories lex(clara, name, human). lex(maria, name, human). lex(juan, name, human). lex(barcelona, name, locative).

Nouns’ semantic categories lex(libro, n, K, libro(K), inanimate). lex(hombre, n, K, hombre(K), human). lex(profesor, n, K, profesor(K), human). lex(perro, n, K, perro(K), animate). lex(gato, n, K, gato(K), animate). lex(camino, n, K, camino(K), locative).

Verbs’ semantic categories lex(piensa, v, S, [arg(en, O)], pensar_en(S, O), sem(human, [X])). lex(habla, v, S, [arg(de, O), arg(con, O 1)], comunica(S, O, O 1), sem(human, [X, human])). lex(está, v, S, [arg(en, O)], locative(S, O), sem(X, [locative])). lex(lee, v, S, [arg(null, O)], leer(S, O), sem(human, [inanimate])).

Grammar 5: semantic constraints v(S, A, F, sem) [Word], {lex(Word, v, S, A, F, sem)}. name(Word, sem) [Word], {lex(Word, name, sem)}. n(Word, sem) [Word], {lex(Word, n, sem)}. np(K, S 2, F, sem) det(K, S 1, S 2, F), n(K, S 1, sem). np(K, F, F, sem) name(K, sem).

Grammar 5: semantic constraints compl([ ], S, S, [ ]) [ ]. compl([arg(X, K) | Y], S 1, S, [sem | sem 2]) p(X), np(K, S 2, S, sem), compl(Y, S 1, S 2, sem 2). s(S) np(K, S 2, S, sem 1), v(K, X, S 1, sem(sem 1, list -sem)), compl(X, S 1, S 2, [list-sem]).