lexical rules may themselves be meaning-bearing "constructions" of several morphemes at a level smaller that a phrase.
1. utterance token
2. construction instance, wordform, morph
3. construction class, lexical item, morpheme
4. schematic construction, lexical rule, derivation pattern of morphemes (abstraction over a class, reducing it to essential elements)
5. relations of constructions (central, peripheral)
6. hierarchical lexicogrammar
2b. referring act or speech act segment of building a shared information state
For Austronesian,
study morpholexical constructions of stems in usage as different parts of speech
Goldberg + "Construction as Categorization of Language"
"Exactly which categories are formed is determined to a great extent by the usefulness of a potential category in predicting how the environment will behave. ... Function plays an obvious and important role in determining what is categorized and affords similar predictive value to the category. ... When it comes to language, clearly the relevant predictive tasks are to predict meaning, given the form (comprehension) and to predict form, given meaning (production). Thus given the fact that constructions relate critical aspects of form and meaning, their acquisition is well-motivated. Accordingly, earlier work has emphasized the role of prediction in language learning, and has shown that constructions are good predictors of overall sentence meaning, useful for conveying “who did what to whom” (Bencini & Goldberg, 2000; Goldberg, Casenhiser, & Sethuraman, 2005)."
Thomas of Erfurt on modes of signifying
Sunday, August 31, 2008
Saturday, August 30, 2008
Relations and roles in action verbs
Language speakers or signers use a verb to pick out a constraint in a scene (or a less concrete situation, often by metaphor) where something is happening. The verb classifies the scene with as an event type, and the verb-constraint is the relation of the Before subscene-type to an After subscene type. Attunement to the verb-constraint allows a speaker-hearer who recognizes (and perhaps expresses with) the verb to construct a shared information state with a number of expected participants, some of which are obligatory in the syntax. What is happening in the action scene is there is a relevant shift from a Before configuration of participants to an After configuration. There may be many different relations picked out in the scene (to a greater or lesser extent), but the most important one is the verb-constraint linking Before and After. It allows the users of the verb-containing utterance to acquire information about the action scene in time, that the action started presupposing the participants in the Before subscene-type and resulted in a situation of an After subscene-type with the participants in a new configuation. A clausal expression predicates that the verb-constraint obtains in the action scene, and creates conditions of satisfaction for a successful communicative act. The receiver of the utterance, if attuned to the verb-constraint, can understand those conditions of satisfaction which can affect their cognition via a shared information state and subsequently their perception and action in the world.
Other relations in the action scene are between the participants. However, certain relations are made explicit by the form of the utterance, by the relation of the participant argument (typically a referring phrase, an NP) to the head verb. The most important (and thus least oblique) argument role defines a relation to the action mentioned by the head verb called a thematic role of Agent. This is in a world where relevant action is performed by other humans, or animals that behave with intentionality and volition similar to humans. A more general thematic role that does not require intention is Effector. Both Agents and Effectors pick out the Initiator of an action occurring. In the Stanford CSLI Verb Semantic Ontology, the two are sometimes not distinguished, and assigned the thematic Role of EFNT, which I guess means Effector-Agent but could simple be pronounced Initiator.
A prototypical type of action scene involves an Agent manipulating an object in the scene. That object may take the role of Patient or Theme. A Patient is structurally changed by the Action, while a Theme may be moved, said or experienced but does not undergo structural change.
Sowa's IRMPACO ontology in Appendix B of his book Knowledge Representation [Sowa 2000] sets out 18 thematic roles, classified according to four kinds that are reminiscent of Aristotle's four aitiae (singluar aitia), traditionally translated as causes (effective, material, formal, ultimate). In an orthogonal dimension, they are classified as Action, Process, Transfer, Spatial, Temporal and Ambient.
It is worth analyzing those 18 roles further, in relation to Goldberg's analysis of constructions (and FrameNet's semantic frames), but that is the subject of a different posting. This a way of exploring verbs and their arguments, and can be contributed to the Stanford CSLI Semantic Verb Ontology. That resource already considers WordNet, I would want to explore mappings to COBUILD and FrameNet.
Other relations in the action scene are between the participants. However, certain relations are made explicit by the form of the utterance, by the relation of the participant argument (typically a referring phrase, an NP) to the head verb. The most important (and thus least oblique) argument role defines a relation to the action mentioned by the head verb called a thematic role of Agent. This is in a world where relevant action is performed by other humans, or animals that behave with intentionality and volition similar to humans. A more general thematic role that does not require intention is Effector. Both Agents and Effectors pick out the Initiator of an action occurring. In the Stanford CSLI Verb Semantic Ontology, the two are sometimes not distinguished, and assigned the thematic Role of EFNT, which I guess means Effector-Agent but could simple be pronounced Initiator.
A prototypical type of action scene involves an Agent manipulating an object in the scene. That object may take the role of Patient or Theme. A Patient is structurally changed by the Action, while a Theme may be moved, said or experienced but does not undergo structural change.
Sowa's IRMPACO ontology in Appendix B of his book Knowledge Representation [Sowa 2000] sets out 18 thematic roles, classified according to four kinds that are reminiscent of Aristotle's four aitiae (singluar aitia), traditionally translated as causes (effective, material, formal, ultimate). In an orthogonal dimension, they are classified as Action, Process, Transfer, Spatial, Temporal and Ambient.
It is worth analyzing those 18 roles further, in relation to Goldberg's analysis of constructions (and FrameNet's semantic frames), but that is the subject of a different posting. This a way of exploring verbs and their arguments, and can be contributed to the Stanford CSLI Semantic Verb Ontology. That resource already considers WordNet, I would want to explore mappings to COBUILD and FrameNet.
Friday, August 29, 2008
prototypical scenes of manipulative activity in grammar, lexicon and concept acquisition
I am reading Goldberg 1995 on Constructions, and she cites Slobin's observation that "children's first use of grammatical markings are applied to 'prototypical scenes'," and that transitivity ("the object of a verb") begins with "manipulative activity scenes".
We can take this thinking beyond grammar, to the cognitive grounding of object recognition. [Note that object in this context refers to an ontological object, Sowa's Independent Physical Continuant; not as in the earlier use of the word to mean the grammatical function of Direct Object] Of course children already have visual object recognition, but that may be primarily in the procedural circuit (the child's own manipulations), not yet define the the discrete schemes of semantic memory since they are still acquiring their first language. To some extent there is likely a discrete conceptual scheme of perceived objects (seen and touched, or for food tasted and smelled), but this is just starting to be mapped to a verbal scheme with phonological forms triggering lexical constructions. In a sense, the lexical constructions must be in the head, but the relationship of semantic types with the phonological form is almost completely conventional (but see my previous remarks on the motivated relationship between sign language morphology and semantics).
We may be realist about the semantic types (they are "out there" in the extra-mental world), but even if we aren't (they might be Saussure's signifieds) the conventional relationship is surely out there in the social world of conventions.
During early stages of concept then word acquisition, what is the underlying (to some extent innate) scheme infrastructure or type system and what are the culturally shaped concept or word schemas? Scheme refers to a more or less extensive scheme of individuation with situations, situation-types, object-types, relations and parametrized states of affairs (infons). A schema refers to a specific complex type that is instantiated during a particular occurrent cognitive process or utterance. Characterizing this needs to consider what happens with otherwise cognitively-normal children who can acquire concept schemas, but have disability barriers to proceeding to word schemas, i.e. language-isolated deaf children. An account adequate for normal (evolutionarily speaking) and variant humans also needs to make sense with other hominids in evolutionary time and their conceptual schemas or counterpart. Certainly all hominids have sophisticated visual object recognition, and to some extent classify visual object types (according to use value?). What is the difference with a human scheme for discrete semantic memory, verbal or non-verbal?
If children begin their grammar acquisition as observed, then "object markers is accusative languages and subject markers in ergative languages are first applied to the arguments of verbs involving direct physical action, e.g. give, grab, take, hit, and not to those of verbs such as say, see, call out."
I might point out that in basal Austronesian languages, many manipulative verbs have a preferred accusative-type transitivity, while certain other experience verbs have a preferred ergative-type transitivity (either patient or locative/dative). [Need to test this intuition with examples.]
Goldberg's objective is to characterize the most senses of common phrasal constructions, and to understand the relationship with other less central senses. Her example is the ditransitive construction in English:
Although this work is about grammatical objects in a transitive construction, I also want to relate it to the metaphysical-cognitive basis of objects in general. It is possible to consider situation-types and particularly events or scenes as prior to the individuation of object-types.
Hominids scanning their visual field bring attentional focus to "what is happening". They may referential actions (like pointing, or naming, or mentioning the noun of an object type, or structure noun phrases) to achieve a shared information state with an interlocutor hominid. The salience of this reference often depends of change of position or location (during manipulation, for example) or change of state. Individuating a perceived thing as a continuant object is premised on scenes where something or nothing of interest is happening. And proceeding beyond NP's to clauses, happening scenes or events get clasified by the verbal relations, where the subscene before the individuated action is a situation-type that stands in the mentioned relation to the same subscene after the action. Picking out such a dimension of an event happening is precisely what it means to mention an action with a verb.
So a verb-defined action-relation that is not just a simple relation (like those of propositions and some modifiers) but a constraint between a before-subscene and an after-subscene. The constraint is a meaning bearing relation, that among other things allows the speaker-hearer to pick out a change in the configuration of participants in the action scene. The participants as objects (or persons or creatures or stuff) are involved in the scene in a particular participant role. At the meta-level these participant roles can be considered at a fine granularity (appropriate to a specific verb), a medium granularity (in a semantic frame, possibly a role unique to only a few frames), or a the coarse granularity of thematic roles that are visible in syntax, and this may itself divide into positions in obliqueness-ordered argument structure and grammatical functions as subject and/or complements. In some Austronesian languages like Tagalog/Filipino, the ang is both determiner and marker and the ang-marked NP is both a subject and simultaneously a complement with surface case marking.
Not all verbs are action verbs, there are also stative verbs. "Words like remain, bother, appear, and exist are verbs, but do not involve any action." [Kim and Sells 2008] In agglutinative Austronesian, there are many roots which are flexibly verb or noun depending on the affixation, and yet take arguments (scene participants) in either discourse context.
So action verb relations are information-bearing constraints. The prototypical scenes that they describe or manifest speech acts involve constraints like cooccurrence, causation, volitional intention, and non-volitional intentionality (like passive perception). A set of named constraints including these might provide a controlled vocabulary or ontology for describing the meaning contributions of specific senses of a phrasal or lexical construction. What is FrameNet doing in this area?
[Kim and Sells 2008] Jong-Bok Kim and Peter Sells. English Syntax: An Introduction. CSLI, 2008.
We can take this thinking beyond grammar, to the cognitive grounding of object recognition. [Note that object in this context refers to an ontological object, Sowa's Independent Physical Continuant; not as in the earlier use of the word to mean the grammatical function of Direct Object] Of course children already have visual object recognition, but that may be primarily in the procedural circuit (the child's own manipulations), not yet define the the discrete schemes of semantic memory since they are still acquiring their first language. To some extent there is likely a discrete conceptual scheme of perceived objects (seen and touched, or for food tasted and smelled), but this is just starting to be mapped to a verbal scheme with phonological forms triggering lexical constructions. In a sense, the lexical constructions must be in the head, but the relationship of semantic types with the phonological form is almost completely conventional (but see my previous remarks on the motivated relationship between sign language morphology and semantics).
We may be realist about the semantic types (they are "out there" in the extra-mental world), but even if we aren't (they might be Saussure's signifieds) the conventional relationship is surely out there in the social world of conventions.
During early stages of concept then word acquisition, what is the underlying (to some extent innate) scheme infrastructure or type system and what are the culturally shaped concept or word schemas? Scheme refers to a more or less extensive scheme of individuation with situations, situation-types, object-types, relations and parametrized states of affairs (infons). A schema refers to a specific complex type that is instantiated during a particular occurrent cognitive process or utterance. Characterizing this needs to consider what happens with otherwise cognitively-normal children who can acquire concept schemas, but have disability barriers to proceeding to word schemas, i.e. language-isolated deaf children. An account adequate for normal (evolutionarily speaking) and variant humans also needs to make sense with other hominids in evolutionary time and their conceptual schemas or counterpart. Certainly all hominids have sophisticated visual object recognition, and to some extent classify visual object types (according to use value?). What is the difference with a human scheme for discrete semantic memory, verbal or non-verbal?
If children begin their grammar acquisition as observed, then "object markers is accusative languages and subject markers in ergative languages are first applied to the arguments of verbs involving direct physical action, e.g. give, grab, take, hit, and not to those of verbs such as say, see, call out."
I might point out that in basal Austronesian languages, many manipulative verbs have a preferred accusative-type transitivity, while certain other experience verbs have a preferred ergative-type transitivity (either patient or locative/dative). [Need to test this intuition with examples.]
Goldberg's objective is to characterize the most senses of common phrasal constructions, and to understand the relationship with other less central senses. Her example is the ditransitive construction in English:
- Central sense: Agent successfully causes recipient to receive patient
- verbs that inherently signify acts of giving: give pass hand serve feed ...
- verbs of instantaneous cause of ballistic motion: throw toss slap kick poke fling shoot ...
- verbs of __: bring take ...
- Conditions of Satisfaction imply that agent causes recipient to receive patient
- verbs of __: guarantee promise owe ...
- Agent causes patient not to receive patient
- verbs of __: refuse, deny ...
- A acts to cause R to receive P at some future point in time
- verbs of __: leave bequeath allocate reserve grant ...
- A enables R to receive P
- verbs of __: permit allow ...
- A intends to cause R to receive P
- verbs of __: bake make build cook sew knit ...
- verbs of __: get grab win earn ...
Although this work is about grammatical objects in a transitive construction, I also want to relate it to the metaphysical-cognitive basis of objects in general. It is possible to consider situation-types and particularly events or scenes as prior to the individuation of object-types.
Hominids scanning their visual field bring attentional focus to "what is happening". They may referential actions (like pointing, or naming, or mentioning the noun of an object type, or structure noun phrases) to achieve a shared information state with an interlocutor hominid. The salience of this reference often depends of change of position or location (during manipulation, for example) or change of state. Individuating a perceived thing as a continuant object is premised on scenes where something or nothing of interest is happening. And proceeding beyond NP's to clauses, happening scenes or events get clasified by the verbal relations, where the subscene before the individuated action is a situation-type that stands in the mentioned relation to the same subscene after the action. Picking out such a dimension of an event happening is precisely what it means to mention an action with a verb.
So a verb-defined action-relation that is not just a simple relation (like those of propositions and some modifiers) but a constraint between a before-subscene and an after-subscene. The constraint is a meaning bearing relation, that among other things allows the speaker-hearer to pick out a change in the configuration of participants in the action scene. The participants as objects (or persons or creatures or stuff) are involved in the scene in a particular participant role. At the meta-level these participant roles can be considered at a fine granularity (appropriate to a specific verb), a medium granularity (in a semantic frame, possibly a role unique to only a few frames), or a the coarse granularity of thematic roles that are visible in syntax, and this may itself divide into positions in obliqueness-ordered argument structure and grammatical functions as subject and/or complements. In some Austronesian languages like Tagalog/Filipino, the ang is both determiner and marker and the ang-marked NP is both a subject and simultaneously a complement with surface case marking.
Not all verbs are action verbs, there are also stative verbs. "Words like remain, bother, appear, and exist are verbs, but do not involve any action." [Kim and Sells 2008] In agglutinative Austronesian, there are many roots which are flexibly verb or noun depending on the affixation, and yet take arguments (scene participants) in either discourse context.
So action verb relations are information-bearing constraints. The prototypical scenes that they describe or manifest speech acts involve constraints like cooccurrence, causation, volitional intention, and non-volitional intentionality (like passive perception). A set of named constraints including these might provide a controlled vocabulary or ontology for describing the meaning contributions of specific senses of a phrasal or lexical construction. What is FrameNet doing in this area?
[Kim and Sells 2008] Jong-Bok Kim and Peter Sells. English Syntax: An Introduction. CSLI, 2008.
sign language phonology-semantics is NOT unmotivated
I am reading Kim and Sells 2008 [1] (http://web.khu.ac.kr/~jongbok/research/eng-syn-draft.pdf ), an HPSG undergraduate text on the syntax of English. They cite Saussure:
They go on to discuss another principle:
They cite Chomsky's notion of grammatical competence an internalized: "Competence refers to speakers’ internalized knowledge of their language." They side with Chomsky's Cartesian "meaning-in-the-head" position. But can't meaning be at least partly in the world? Taking a Gibsonian ecological psychology stance (or the position of Clark and Chalmers), can't we use the external perceived word as a major memory support structure? This might be a bit limited for sounds of speech, but there is more opportunities with the spaces of signs. Think of how pronouns in sign language are created in discourse. Signers establish a place in space, then refer back to it by pointing. Or they might count of a set of individuals on their fingers, then come back to each finger. So knowledge, including rules, may not be completely in the head. This reminds me of Barwise's realism about types and "constraints" (relations between two situation-types) in his relational theory of meaning. If situation-theoretic constraints are realities in the extramental world, then it seems meaning is in the world, and it is attunement to constraints that is (partly) in the head. I say partly, because attunement to language in practice cannot be a "private language," it involves a shared scheme of individuation of linguistic conventions. There may be some innate base for the conventions, the neural correlates of a the bare bones of a linguistic type system, but the cognitive structures get tuned to a speech communities linguistic constraints. Even a cat gets their binocular vision tuned to the regularities of the external world, a proven by monocular cats with no depth vision.
I am sure I will have more commentary to come.
----------------
[1] English Syntax: An Introduction Jong-Bok Kim and Peter Sells
The first well-known property (as emphasized by Saussure 1916) is that there is no motivated relationship between sounds and meanings. This is simply observed in the fact that the same meaning is usually expressed by a different sounding-word in a different language (think of house, maison, casa). For words such as hotdog, desk, dog, bike, hamburger, cranberry, sweetbread, their meanings have nothing to do with their shapes. For example, the word hotdog has no relationship with a dog which is or feels hot. There is just an arbitrary relationship between the word’s sound and its meaning: this relationship is decided by the convention of the community the speakers belong to.However, onomatopoeia shows there can be some link. And for sign language phonology, there is a very extensive link between the spatial and movement aspects of the gestures and the spatial and movement elements of the described situation. There is still a certain arbitrariness to the codified conventions of, let us say, vocabulary, but there is a significant relationship which plays a role in interpretation (and mnemonics).
They go on to discuss another principle:
The second important feature of language, and one more central to syntax, is that language makes infinite use of finite set of rules or principles, the observation of which led the development of generative linguistics in the 20th century (cf. Chomsky 1965). A language is a system for combining its parts in infinitely many ways. One piece of evidence of the system can be observed in word-order restrictions. If a sentence is an arrangement of words and we have 5 words such as man, ball, a, the, and kicked, how many possible combinations can we have from these five words? More importantly, are all of these combinations grammatical sentences? Mathematically, the number of possible combinations of 5 words is 5! (factorial), equalling 120 instances. But among these 120 possible combinations, only 6 form grammatical English sentences:Now at the general level, I think this claim is quite true of sign languages. However, there is a problem in generalizing the idea that rules are strictly combinatorial. This is an artifact of the serializing channel of phonemes, while sign languages have multiple concurrent channels of emic elements, that combine in space-dependent ways. So you can't take factorials to characterize sign language rules.
They cite Chomsky's notion of grammatical competence an internalized: "Competence refers to speakers’ internalized knowledge of their language." They side with Chomsky's Cartesian "meaning-in-the-head" position. But can't meaning be at least partly in the world? Taking a Gibsonian ecological psychology stance (or the position of Clark and Chalmers), can't we use the external perceived word as a major memory support structure? This might be a bit limited for sounds of speech, but there is more opportunities with the spaces of signs. Think of how pronouns in sign language are created in discourse. Signers establish a place in space, then refer back to it by pointing. Or they might count of a set of individuals on their fingers, then come back to each finger. So knowledge, including rules, may not be completely in the head. This reminds me of Barwise's realism about types and "constraints" (relations between two situation-types) in his relational theory of meaning. If situation-theoretic constraints are realities in the extramental world, then it seems meaning is in the world, and it is attunement to constraints that is (partly) in the head. I say partly, because attunement to language in practice cannot be a "private language," it involves a shared scheme of individuation of linguistic conventions. There may be some innate base for the conventions, the neural correlates of a the bare bones of a linguistic type system, but the cognitive structures get tuned to a speech communities linguistic constraints. Even a cat gets their binocular vision tuned to the regularities of the external world, a proven by monocular cats with no depth vision.
I am sure I will have more commentary to come.
----------------
[1] English Syntax: An Introduction Jong-Bok Kim and Peter Sells
Thursday, August 28, 2008
"wanna" is a lexical item, not just a contraction
I was reading The Analytic/Synthetic Distinction (Georges Rey) [REVISED: August 15, 2008] and came across this:
"An ungrammatical sentence like “Bill is the man I wanna take a walk” might suffice on occasion for thought and communication (of “Bill is the man who I want to take a walk,”), but it's a striking fact that speakers of English—even four-year old ones!—nevertheless find it problematic (see Crain and Lillo-Martin 1999)."
"An ungrammatical sentence like “Bill is the man I wanna take a walk” might suffice on occasion for thought and communication (of “Bill is the man who I want to take a walk,”), but it's a striking fact that speakers of English—even four-year old ones!—nevertheless find it problematic (see Crain and Lillo-Martin 1999)."
Thursday, August 21, 2008
event-types, and emic signals in sign phonology
This needs cleaning up, but I figured I should post this in draft form, so I don't lose it.
How do signs pick out place?
Dominant-Hand emic signal
Compare with speech utterances:
Syllables
Prosody
Humanly relevant scene
Typical Humanly-Relevant Scene
Agent1
How do signs pick out place?
Dominant-Hand emic signal
- shape
- orientation
- shape
- orientation
- Start Config
- Motion
- path
- manner
- repeat
- End Config
- Face
- Eyes
- Eyebrows
- Lips
- Shoulders
Compare with speech utterances:
Syllables
- Initial
- Rime
- nucleus
- coda
Prosody
Humanly relevant scene
- a token situation accessible to physical perception
- a way of classifying scenes and other locations as located in Place and Time
- one participant + speaker
- recognize object (e.g. Proper Name, or deictic Mama)
- identity condition, a later ocurrent percept is the same object
- discrete memory (as opposed to procedural memory)
- tracked or inferred continuant
- recognize happening (precursor to Action)
- precondition situation
- result situation
- situation properties (e.g. manner)
- Refer to an object of a type
- speaker and hearer share schema for that type in their schemes of individuation
- types are made explicit in the scene, such as via an utterance (noun phrase)
- conditions of success -- may fail to refer successfully
- may be performed with eye gaze, head glance, deictic pointing (e.g. hand), or uttering a common noun, or a pronoun
- anaphoric resolution; successive references may refer to the same index in discourse
- modifiers of a head noun help resolve which of several possible objects are the target of referring
- Request Action
- Conditions of Satisfaction: hearer acts so that mentioned action results obtain in the world
- Assert Action
- Conditions of Satisfaction: the world matches the mentioned action result
- Assert State
- Conditions of Satisfaction:: the world or world objects have properties that match the mentioned state
- Promise Action
- Conditions of Satisfaction:: the speaker takes action so that the mentioned action results obtain in the world
Typical Humanly-Relevant Scene
Agent1
- locomote
- move body-part
- manipulate object.graspable
- perceive
- display
- eye gaze
- body movement
- facial expression
- smile, frown
- gesture
- vocal
- laughter, crying
- cry out for attention
- utterance
- speech
- sign language utterance
- cause
- experience
- receive
- affected
- beneficiary
- move
- path
- start
- end
- have-state
- instrument
- channel-of-motion
- Place2 (part of Place1)
- Time.Duration2
- Time.Point1
- Precondition situation
- Result situation
- Constraint
- information-bearing
- causal constraint
Monday, August 18, 2008
Limits of Formal Ontology
I am thinking of volunteering to help with the Stanford-CSLI Semantic Verb Ontology, which would involve studying Prolog.
A passage in Sowa's Knowledge Representation book Appendix B reminded me of how heroically radical some of the assumptions of Formal Ontology are.
This seems to be committed to the tradition of compositional semantics from lexical tokens. In construction grammar, the constructions themselves contribute meaning, not just the lexical items. Perhaps it is possible to rescue a multi-level compositionality, as in the type hierarchy of HPSG-style CxG (in Sag's draft), but again maybe not.
Also, there seems to be a tradition strong autonomy of the grammar (the syntax, and also the structural semantics) from the lexicon (with its lexical semantics). On the other hand, systemic functional theories have long emphasized an integrated lexicogrammar as the field where functional choices are made.
A passage in Sowa's Knowledge Representation book Appendix B reminded me of how heroically radical some of the assumptions of Formal Ontology are.
An informal ontology may be specified by a catalog of types that are either undefined or defined only by statements in a natural language. A formal ontology is specified by a collection of names for concept and relation types organized in a partial ordering by the type-subtype relation.This fixes a set of names, in a lexical ontology, the set of lexical items in the shared verbal lexicon of the speech/sign community. More questionable is that it fixes the set of types, where the cognitive reality may be dynamically changing family resemblances of language games.
This seems to be committed to the tradition of compositional semantics from lexical tokens. In construction grammar, the constructions themselves contribute meaning, not just the lexical items. Perhaps it is possible to rescue a multi-level compositionality, as in the type hierarchy of HPSG-style CxG (in Sag's draft), but again maybe not.
Also, there seems to be a tradition strong autonomy of the grammar (the syntax, and also the structural semantics) from the lexicon (with its lexical semantics). On the other hand, systemic functional theories have long emphasized an integrated lexicogrammar as the field where functional choices are made.
Sunday, August 17, 2008
Verb Semantic Ontology
On the Conceptual Graphs List, John Sowa mentioned this project that started at IBM and involved VivoMind, including Arun Majumdar.
Some of the people involved: Beth Levin and John F. Sowa, Bonnie J. Dorr (Lexical Conceptual Structure), Martha Palmer (VerbNet), Timothy Chklovski (VerbOcean ), and, Charles Fillmore ( FrameNet ).
Some of the people involved: Beth Levin and John F. Sowa, Bonnie J. Dorr (Lexical Conceptual Structure), Martha Palmer (VerbNet), Timothy Chklovski (VerbOcean ), and, Charles Fillmore ( FrameNet ).
The boundary between words and nonwords, the Borgmann project to list all words of English
I watched a lecture by Chris Cole on Google Tech Talks:
The Borgmann Project: Listing all the Words in English
http://www.youtube.com/watch?v=kU_CiErwkFA#
Here is a critical blog post on it:
http://cqs.livejournal.com/48862.html
By eternally stressed semanticist Lance Nathan, at U Penn Linguistics.
I am more sympathetic to the project. I like his insight that the "list" will be a process that finds a stable solution from top down (generating rules to get probabilities) and bottom up (occurrences in a corpus).
The Borgmann Project: Listing all the Words in English
http://www.youtube.com/watch?v=kU_CiErwkFA#
Here is a critical blog post on it:
http://cqs.livejournal.com/48862.html
By eternally stressed semanticist Lance Nathan, at U Penn Linguistics.
I am more sympathetic to the project. I like his insight that the "list" will be a process that finds a stable solution from top down (generating rules to get probabilities) and bottom up (occurrences in a corpus).
Subscribe to:
Posts (Atom)