Guiding Assumptions of Cognitive Grammar

1
Guiding Assumptions of Cognitive Grammar

• Gui Shichun (baed on Langacker)

2

• SymbolizationLangackers thinking departs in
  some ways from the classic Saussurean paradigm.
• The arbitrary character of linguistic signs is
  easily overstated. Any polymorphemic linguistic
  sign is nonarbitrary to the extent that it is
  analyzable. (staple vs stapler)
• The conception of language as symbolic in nature
  extends beyond lexicon to grammar. Grammar is
  inherently symbolic and hence meaningful.

3

• CognitionLanguage is an integral part of human
  cognition.
• An account of linguistic structure should
  articulate with what is known about cognitive
  processing in general.
• Language has appeared special and unassimilable
  to broader psychological phenomena mainly because
  linguists have insisted on analyzing it in an in-
  appropriate and highly unnatural fashion.
• NaturalnessNaturalness is an essential criterion
  for evaluating linguistic theories and analyses.

4

• A description is natural to the extent that it
  deals with data in their own terms, with full
  regard for the richness, subtlety, and complexity
  characteristic of linguistic phenomena.
• A description is unnatural, or artifactual, to
  the extent that it deals with data in a way that
  does violence to their intrinsic organization,
  however convenient this may be for the analyst.
• Discretenessthe issue of discreteness has four
  dimensions.
• Simple categorical judgment. A simple /- value
  or yes/no answer is not always sufficient in
  specifying whether a linguistic structure has a
  certain property, belongs to a particular
  category, or participates in a given
  relationship. These conditions are often matters
  of degree, and we must devise some means of
  accommodating the complete range of
  possibilities. (e.g. drive a nail vs drive a car,
  ring as a sound, as a boxing arena, and as a
  piece of jewelry?)

5

• Models for categorization. The choice between the
  standard criterial-attribute model of linguistic
  categorization and a conception based on
  prototypes.
• The criterial-attribute model characterized a
  class by means of a list of defining features it
  requires that every member of the class fully
  possess every property on that list, and that no
  nonmember possess all of the listed properties.
  Class membership is thus an all-or-nothing
  affair.
• The prototype model does not require that every
  member of a category possess a given feature, but
  it encourages the fullest possible
  characterization of prototypical instances.
  Finally, the prototype model allows an entity to
  be assimilated to a category if a person finds
  any plausible rationale for relating it to
  prototypical members.

6

• Dichotomous organization. Langacker regards the
  following rigid dichotomies as false synchrony
  vs. diachrony, competence vs. performance,
  grammar vs. lexicon, morphology vs. syntax,
  semantics vs. pragmatics, rule vs. analogy,
  grammatical vs. ungrammatical sentences,. One
  way to produce a false dichotomy is to focus
  solely on representative examples from the two
  extremes of a continuum by overlooking
  intermediate cases, one readily observes discrete
  classes with sharply contrasting properties.
  (e.g. content words, such as giraffe,
  encyclopedia, vs. functional forms such as ing,
  of, be) the differences are of course striking.
  But content words vary enormously in concreteness
  of meaning (kickgttalkgtthinkgtlivegtexist) and also
  in semantic specificity (giraffegtmammalgt
  animalgtorganismgtthing)

7

• Integrated systems. Most linguistic units are
  highly integrated structural complexes, or
  systems, which are more than just the sum of
  their recognizable parts. It is nevertheless
  common for linguists to impose a componential
  analysis on these systems , representing them an
  unordered bundles of discrete features. (e.g. the
  concept UNCLE is resolvable into the semantic
  components MALE, COLLATERAL, ASCENDING
  GENERATION.

8

• Substance. Because language is abstract and
  complicated, its characterization demands a
  multitude of abstract notion. The theoretical
  constructs have got to be substantive (real), but
  there is no reliable, recognized authority to
  appeal to in determining whether a linguistic
  construct is substantive.
• All linguists employ traditional concepts (e.g.
  noun, subject, modifier, subordination) that few
  if any can rigorously defined. This leads to the
  rise of structuralism. (e.g. a noun is defined as
  the name of a person, place, or thing, but
  blue is a blue tie is considered as an
  adjective, although it is the name of a color.)

9

• If our understanding of a notion is only
  preliminary and superficial, we must be wary of
  strong conclusions that ultimately depend on
  highly specific assumptions about that notion.
  One example is the use of idioms by generative
  linguists to justify movement rule.
• a. Headway seems to have been steadily made.
• b. ? PAST steadily make headway PRES seem
• To make headway is an idiom which is listed
  in the lexicon as a fixed, unanalyzable
  expression. The deep structures in (b), where ?
  indicates an unspecified subject, are therefore
  required to account for the surface forms in (a).
  But the assumption that idiomatic verb-object
  constructions must always be base-generated
  together is false. (Grimshaw,1978)

10

• (a) We didnt make the amount of headway that was
  expected of us.
• (b) Unfortunately, we made what the president
  considered to be insufficient headway on that
  problem.
• Complexity. Investigators tend to lose sight of
  the actual complexity of their subject matter.
  Idealizations that significantly distort the
  object of investigation are pernicious, and their
  avoidance is critical to the goal of naturalness
  in linguistic analysis. The sharp distinction
  made between lexical and grammatical morphemes is
  an example. Lexical morphemes are clearly
  meaningful, whereas grammatical morphemes are
  often less obvious. So grammatical morphemes are
  often ignored for purposes of grammatical
  analysis.

11

• (a) The Chinese invented gunpowder.
• (b) Gunpowder was invented by the Chinese.
• The two sentences are different in terms of
  information structure. In (a), Chinese is new
  information, as an answer to the question Who
  invented gunpowder? where in (b) gunpowder is
  new information, as an answer to the question
  What was invented by the Chinese. The passive
  voice is a means of reversing the normal order of
  agentive and affected elements. Therefore,
  the original assumption that grammatical
  morphemes and syntactic relationship are
  nonsemantic in character, or that syntax is
  autonomous, is not true.