300 likes | 493 Views
Word Grammar and other cognitive theories. Richard Hudson Budapest March 2012. Cognitive linguistics. Cognitive theories of grammar. CgG. CnG. WG. Shared assumption.
E N D
Word Grammar and other cognitive theories Richard Hudson Budapest March 2012
Cognitive theories of grammar CgG CnG WG
Shared assumption • 'the formal structures of language are studied not as if they were autonomous, but as reflections of general conceptual organisation, categorization principles, processing mechanisms and experiential and environmental influences' • Geeraerts and Cuyckens 2007:3
The Cognitive Principle • 'Knowledge of language is knowledge' • Goldberg 1995:5 • Contrast Modularity • Language is a separate 'module' of the mind. • Let's call this the Cognitive Principle.
Different notations • CgG • e.g. Langacker 2007 • CnG • e.g. Croft 2007, Goldberg 1995 • WG • e.g. Hudson 1980, 1990, 2007, 2010
CnG: Heather sings. (Croft 2007: 476)
WG: Heather sings. singer semantics singing Heather meaning meaning subject syntax Heather sings. • No 'symbolic units'. • Just a network of related concepts.
CgG: (the) table near (the) door (Langacker 2007: 442)
WG: the table near the door landmark table near door position door near door meaning comp comp comp adjunct the table door near the • Just words and other concepts in a network.
Some agreements • grammar-lexicon continuum • no separate lexicon • language is learned from experience (usage) • not innate and 'triggered' • network organisation of language • but what are the nodes?
Some disagreements • Does language consist of symbols? • CgG, CnG: yes WG: no • Is morphology independent of syntax? • CgG, CnG: no WG: yes • What is syntactic structure like? • CgG, CnG: phrases WG: dependencies
Is language 100% symbolic? • "…the pivotal claim of Cognitive Grammar that all valid grammatical constructs have a conceptual characterization" • (Langacker 2007:422) • But: "The CG claim that basic grammatical classes can be characterized semantically … applies to a limited set of categories … • contrast "… idiosyncratic classes … Semantically, the members of such a class may be totally arbitrary." (ibid: 439)
… and Construction Grammar • "In Construction Grammar, the basic linguistic units are symbolic and are organized as symbolic units" • Croft 2007:473 • But: Some constructions have no meaning, e.g. Subject-Auxiliary Inversion • ibid: 484 • So some units are not symbolic.
Against symbols • Meanings and forms do not match. • Some forms or classes have no meaning • e.g. 'irregular verb' • Some 'meanings' cannot be expressed • e.g. 'sibling', German fahren • Some forms express complex meanings • e.g. verbs like GIVE, LEND, MAKE …
CnG: the Benefactive-Ditransitive construction (Goldberg 1995: 77)
The Goldberg analysis • Semantics and syntax are totally in step: • one verb, e.g. give, lend • one predicate, e.g. CAUSE-RECEIVE • three arguments for one predicate: • agent • recipient • patient
But: John lent Mary his car. • = 'John caused Mary to receive his car' • two predicates, with separate arguments: • Pred1: John caused Pred2 • Pred2: Mary received his car. • Pred1 is an action (John lent … at noon) • Pred2 is a state (John lent … for two days)
Semantics and syntax are independent • So we need an analysis which allows semantics and syntax not to be in step. • e.g. 'Benefactive ditransitive construction' • John made Mary a cake. • Syntax: one verb, three dependents • Semantics: at least two predicates: • Pred1: John made a cake in order for Pred2 • Pred2: Mary had the cake.
WG: the Benefactive-Ditransitive construction 'is-a' subject • No constructions. • Just words and other concepts • Default inheritance applies to words. verb • object • transitive 'rec' result having • • ind obj • ditransitive purpose beneficiary • • • ind obj benefactive ditransitive •
Morphology is independent of syntax too • Homonyms: two words, one morph • e.g. STICKn or STICKv = {stick} • learner must recognise {stick} before STICK • Clitics: two words, one morph • e.g. YOU + BE:pres = {your} = /jɔ:/ • Fusion: many functions, one morph • e.g. Latin: present, singular, 1st-person = {o}
The architecture of language in WG semantics meaning syntax realisation morphology realisation graphology phonology
Syntactic structures • "… a construction … is made up of parts, and those parts are themselves independent constructions." • Croft 2007: 495 • But: "In Cognitive Grammar … grammatical constituency is … variable, nonessential and nonfundamental." • Langacker 2007: 442
Phrase structure in CgG, CnG • Very simple phrase structure • The only relations possiblein syntax are: • part-whole (sub-classified for function) • left-right • A very odd assumption for cognitive linguists • because we easily handle many other relations outside language, e.g. between people.
For example, a kinship network Gretta son brother mother husband Colin me Gaynor brother wife daughter daughter grandson Lucy son Peter
WG syntax • Dependency structure • like school grammar • but much richer • Dependencies: • are asymmetrical • link single words • can be sub-classified eg. as 'subject', 'adjunct'
A simple example subject adjunct object adjunct English visitors generally like Budapest
A richer example comp pred pred pred extractee subject Where do they tend to stay? s x comp
Conclusion • Language-knowledge is just knowledge. • It's a network of nodes (not of boxes). • Semantics is independent of syntax. • So is morphology. • Syntax is a network of dependency relations among words.
Thank you • This talk can be downloaded: www.phon.ucl.ac.uk/home/dick/talks.htm • More on Word Grammar: www.phon.ucl.ac.uk/home/dick/wg.htm