Optimality Theory Lexical Semantics Tandem workshop on Optimality

  • Slides: 24
Download presentation
Optimality Theory Lexical Semantics Tandem workshop on Optimality Theory in language and geometric approaches

Optimality Theory Lexical Semantics Tandem workshop on Optimality Theory in language and geometric approaches to language

Overview � Overview OT & lexical semantics � Consequences of OT view on lexical

Overview � Overview OT & lexical semantics � Consequences of OT view on lexical semantics � Content words and OT � Uni- versus bidirectional OT � Experimental testing � Conclusions

OT lexical semantics Previous work: • Fong (2003): use of already in Colloquial Singaporean

OT lexical semantics Previous work: • Fong (2003): use of already in Colloquial Singaporean English and Standard English • Zwarts (2004): interpretation of preposition (a)round • Zeevat (2002): various discourse markers • Zwarts (2008): production of prepositions • Hogeweg (2009): discourse marker wel

OT lexical semantics Lexicon Production Word 1: {f 1, f 2, f 3, f

OT lexical semantics Lexicon Production Word 1: {f 1, f 2, f 3, f 4} Word 2: {f 1, f 5, f 6, f 8, f 9} Word 3: {f 3, f 9, f 10, f 11} Word 4: {f 11, f 12, f 13} Word 5: {f 15} Input: meaning/intention (set of features). For example {f 1, f 2, f 9} MARK FAITH DEP word 2: {f 1, f 5, f 6, f 8, f 9} ** *** word 3: {f 3, f 9, f 10, f 11} ** *** word 4: {f 11, f 12, f 13} *** word 5: {f 15} *** * word 1: {f 1, f 2, f 3, f 4} Interpretation Input word from lexicon (set of features) + context. For example: word 4 {f 11, f 12, f 13}. {f 1, f 2, f 3, f 4, f 5} {f 1, f 12, f 13} {f 11, f 12, f 13} {f 12, f 13} {} FIT FAITH *INVENT ***** * * ***

Consequences of OT view on LS � There is no strict relation between word

Consequences of OT view on LS � There is no strict relation between word and meaning but a meaning is the output of the process that takes a word as an input or the input to the process has a word as an output � Whether a concept is labeled by a particular word does not only depend on the stored information for that word but also on the stored information of competing words, that is, competition is important � The meaning of words is overspecified in the lexicon

Consequences: overspecification � Most current theories of lexical semantics argue that lexical representations are

Consequences: overspecification � Most current theories of lexical semantics argue that lexical representations are underspecified and can be strengthened by contextual information (e. g. Reyle 1993, Pustejovsky 1995, Blutner 1998, 2004) � The previous studies (except for the original analysis in Zwarts 2008) assume an overspecified lexical representation, which can be weakened by the context

OT and the semantics of content words �An overspecified lexical representation includes what is

OT and the semantics of content words �An overspecified lexical representation includes what is usually considered conceptual, encyclopedic or commonsense knowledge �For example, lexical knowledge specifies that an apple has a stem, pulp, a peel and that apples can be red or green: {pulp, stem, peel, red ∨ green}

OT and the semantics of content words Lolly pop-frame (based on Petersen 2007) green

OT and the semantics of content words Lolly pop-frame (based on Petersen 2007) green R LO O C body DY O B SHAPE TAS TE Lolly pop R O OL ST ICK stick round apple brown C SH APE long λx[lolly pop(x) ∧ body of (body, x) ∧ color of (green, body) ∧ shape of(round, body) ∧ taste of(apple, body) ∧ etc. ] λx∃y[lolly pop(x) ∧ body of (y, x) ∧ color of (green, body) ∧ shape of (round, bod taste of(apple, body) ∧ etc. ]

OT and the semantics of content words � Hogeweg (submitted): � Ranking of faithfulness

OT and the semantics of content words � Hogeweg (submitted): � Ranking of faithfulness constraints pertaining to attributes and values � Interacting with well-known semantic � Selective binding (Pustejovsky 1995) If α is of type <a, a>, β is of type b, and the qualia structure of β, QSβ, has quale, q of type a, then αβ is of type b, where {αβ} = β ∩ α(qβ) � Non-vacuity principles: principle (Kamp & Partee 1995) In any given context, try to interpret any predicate so that both its positive and negative extension are non-empty’ � Fit (e. g. Zwarts 2005)

OT and the semantics of content words � Well known constraint FIT which penalizes

OT and the semantics of content words � Well known constraint FIT which penalizes candidates with inherent contradictions § § § When there are two or more values for the same attribute, When a value or attribute is of the wrong type as determined by a type hierarchy (for example the value warm for the attribute color for the type smell) When a value or an attribute is in incongruence with an inherited value

OT and the semantics of content words Interpretation lilac apple Input: λx[apple(x) ∧ peel

OT and the semantics of content words Interpretation lilac apple Input: λx[apple(x) ∧ peel FIT NVP *NB FAITH FAITH of (peel, x) ∧ color of (red ∃x[color of ∃x[color of ∨ green, peel) ∧ stem of of(x, (red, (green, of (brown, (stem, x) ∧ color of peel)] peel) (x, stem)] stem) (brown, stem) ∧ etc. ] + λx[lilac (x)] y = apple (peel = red, stem ** * * is brown) y = peel * y = stem (peel = red) y = peel and peel = red * * * *

OT and the semantics of content words Lexicon Production Word 1: {f 1, f

OT and the semantics of content words Lexicon Production Word 1: {f 1, f 2, f 3, f 4} Word 2: {f 1, f 5, f 6, f 8, f 9} Word 3: {f 3, f 9, f 10, f 11} Word 4: {f 11, f 12, f 13} Word 5: {f 15} Input: meaning/intention (set of features). For example {f 1, f 2, f 9} MARK FAITH DEP word 2: {f 1, f 5, f 6, f 8, f 9} ** *** word 3: {f 3, f 9, f 10, f 11} ** *** word 4: {f 11, f 12, f 13} *** word 5: {f 15} *** * word 1: {f 1, f 2, f 3, f 4} Interpretation Input word from lexicon FIT (set of features) + context. For example: word 4 {f 11, f 12, f 13}. {f 1, f 2, f 3, f 4, f 5} {f 1, f 12, f 13} * {f 11, f 12, f 13} {f 12, f 13} {} NVP *NB FAITH ***** * ***

Bidirectional optimization � Aforementioned studies are all cases of unidirectional, interpretive optimization � Blutner

Bidirectional optimization � Aforementioned studies are all cases of unidirectional, interpretive optimization � Blutner (2004) argues that some lexical phenomena can be explained by bidirectional optimization based on a underspecified lexical meaning

Bidirectional optimization �f = form/underspecified representation � m = possible enrichment � f 1>

Bidirectional optimization �f = form/underspecified representation � m = possible enrichment � f 1> f 2 � More features > less features Q I f 1 {fa}, m 1{fa, fb} f 1 {fa}, m 2{fa, fb, fc} * f 2 {fa}, m 1{fa, fb} * *

Bidirectional optimization Not completely comparable: for example interpretation lilac apple ‘lilac’ + ‘apple’ I

Bidirectional optimization Not completely comparable: for example interpretation lilac apple ‘lilac’ + ‘apple’ I y = apple (peel = red, stem is brown) * y = peel y = stem (peel = red) ** y = peel and peel = red ***** Blutner 2009: bidirectional optimization is an offline process that forms language in an evolutionary sense

Metaphor � Metaphor: most extreme examples of the flexibility of word meanings My cousin

Metaphor � Metaphor: most extreme examples of the flexibility of word meanings My cousin is a grey mouse

Psycholinguistic testing � Experiment Fernández (2007): � Subjects presented with spoken sentences: a context

Psycholinguistic testing � Experiment Fernández (2007): � Subjects presented with spoken sentences: a context sentence and a target sentence: Nobody wanted to run against John at school. John was a cheetah. � 60 participants � Set of 22 common nouns (e. g. cheetah) with predictable superordinates (e. g. cat) and distinctive property (fast).

Psycholinguistic testing � The subjects carried out a lexical decision task at 0, 400

Psycholinguistic testing � The subjects carried out a lexical decision task at 0, 400 or 1000 ms after the word recognition point of the critical word. � The target words were presented visually. � If the critical word was cheetah, the target words were for example cat and fast. � Assumption: facilitation relative to an unrelated control is indicative of property activation.

Psycholinguistic testing

Psycholinguistic testing

Psycholinguistic testing � The hypothesis of overspecification leads to a testable prediction for a

Psycholinguistic testing � The hypothesis of overspecification leads to a testable prediction for a similar lexical decision experiment about other semantic phenomena, e. g. coerced nouns and “simple" noun adjective combinations. � Stone lion � Halve appel ‘half an apple’, rotten banana � An overspecification view predicts that upon hearing a noun with an adjective that is in conflict with a highly ranked property of the noun, this property is initially activated and repressed later.

Psycholinguistic testing � An underspecification view would predict that the feature ‘yellow’ is not

Psycholinguistic testing � An underspecification view would predict that the feature ‘yellow’ is not activated because the lexical representation of banana does not include information about its color. � In contrast, this aspect is filled in in the enrichment process based on the context, the immediate context being the adjective rotten in this case.

Psycholinguistic testing �A similar hypothesis holds for the interpretation of stone lion. An overspecification

Psycholinguistic testing �A similar hypothesis holds for the interpretation of stone lion. An overspecification view predicts that interpreting this phrase involves the suppression of features such as ‘carnivore’, ‘runs fast’ and ‘hairy’ because those are not properties of a lion made of stone while they are properties of a (real) lion.

Psychological testing � The experiments could show that “regular interpretation”, (metonymic) type coercion and

Psychological testing � The experiments could show that “regular interpretation”, (metonymic) type coercion and metaphors work similarly, they all involve the suppression of irrelevant features

Conclusions � Previous studies in lexical semantics focused on grammaticalized, functional items � The

Conclusions � Previous studies in lexical semantics focused on grammaticalized, functional items � The studies assume an overspecified lexical representation � Analyses of the interpretation of content words require a more complex representation of meaning and additional constraints � Unidirectional and bidirectional optimization are perhaps not mutually exclusive � The hypothesis of overspecification leads to a testable prediction for a priming experiment