LING 1010 Exam 1 Review Notes

Exam 1 Review Notes

General Exam Information

  • Exam 1 is on Wednesday, February 19th, in ITE C80.
  • The exam is in-person and paper-based. Remember to bring your student ID and a pencil.
  • No electronic devices are allowed during the exam.
  • You may bring one 8.5x11 inch sheet of notes (both sides).
  • The exam consists of 30 multiple-choice questions, each worth 1 point.
  • The duration of the exam is 50 minutes.

Lecture 1: Language Misconceptions

  • Discussed and debunked language misconceptions.
  • One misconception: "Cultures with greater technological sophistication tend to have grammatically richer languages."
  • Evidence against this: The Highlanders of Papua New Guinea, an "experiment of nature."
    • Isolated group with stone-age technology, yet their languages (e.g., Yimas) are as complex as any other known language.
  • Language is not simply a technology but a part of human biology.
  • Compared language to writing, with writing being an invention, implying language is inherent.

Lecture 2: Prescriptive vs. Descriptive Grammar

  • Another misconception: "Careful, formal speech is more grammatically sophisticated than casual, everyday speech."
  • Compared prescriptive and descriptive grammar and their contrasting views.
    • Prescriptive grammar: Sets rules for how language should be used.
    • Descriptive grammar: Describes how language is actually used by native speakers.
  • Examples discussed: Sentence-final prepositions, double negation.
  • Linguists follow descriptive grammar, describing sentences that are grammatical according to native speakers.
  • The Ass Camouflage Construction (ACC):
    • Prescriptively disapproved of.
    • Descriptively interesting due to its systematic grammatical rules, showing a pattern despite being considered incorrect by prescriptive standards.

Lecture 3: Linguistic Competence vs. Performance

  • Distinguished between linguistic competence and linguistic performance.
    • Linguistic competence: A speaker's underlying knowledge of language.
    • Linguistic performance: How a speaker actually uses language (which can be subject to errors).
  • Accessing competence is tricky because linguistic knowledge is largely tacit (unconscious).
  • We cannot directly ask speakers about the rules they follow because the rules are not conscious knowledge.
  • Corpora (large language databases) only provide evidence of performance, not competence.
  • Methods to study competence include:
    • Observations of linguistic performance.
    • Eliciting speakers’ judgments of grammaticality and ambiguity.
  • Wanna contraction as an example:
    • Demonstrated tacit knowledge through grammaticality judgments.
    • Speakers unconsciously know when "want to" can contract to "wanna" based on underlying syntactic structure.

Lecture 4: Creativity and Recursion in Language

  • Human language is fundamentally creative.
  • Languages are infinite, and our linguistic competence enables us to speak and understand novel sentences.
  • Recursion: A linguistic unit containing another unit of the same kind. This property allows for such creativity.
  • Example: Simple recursive rule for generating nominal compounds in English: N→NNN \rightarrow N N.
  • Tree diagrams represent the hierarchical structures created by such rules.
  • Structural ambiguity: Sentences or words with two structures that yield distinct meanings.

Lecture 5: Empiricism vs. Nativism

  • Two different views on how humans obtain knowledge.
  • Empiricists:
    • Emphasize the importance of experience and sensory input.
    • Suggest only general cognitive abilities are required for learning.
  • Nativists:
    • Emphasize innate, domain-specific knowledge required to overcome insufficient input from the environment.
  • Empiricism led to Behaviorism in the 20th century:
    • Focused on studying behavioral responses to environmental stimuli, including human language behavior.
  • Connectionism: A modern Empiricist view.
    • Views artificial neural networks (ANNs) as general-purpose learning devices that can explain language acquisition.
  • Modern Nativism:
    • Acknowledges the necessity of experience in language acquisition.
    • Attributes innate language knowledge to human biology.
  • Noam Chomsky and Universal Grammar:
    • Proposed the concept of Universal Grammar: innate knowledge that allows humans to acquire language despite limited experience.

Lecture 6: Predictions and Responses

  • Predictions of Empiricism:
    • Some people won’t have any language.
    • Some people will take longer to learn it.
    • Different people in a community will have completely different grammars.
    • Different children will go through completely different processes in learning language.
    • Certain languages will be harder for a child to learn.
    • Experience is necessary.
    • Outcomes of language acquisition should bear a relationship to general intelligence.
  • Nativist Response:
    • Universality: All human societies have (always had) language; all children acquire at least one language (aside from pathology).
    • Uniformity:
      • Ease: Language acquisition is generally easy for children.
      • Success: Children generally succeed in acquiring language.
    • Rapidity: Children acquire language relatively quickly.
    • Consistency of stages: Children go through similar stages in language acquisition.
    • Children can succeed in acquiring language even when not exposed to a full language in their primary linguistic data (e.g., creolization).

Lecture 6: Specific Language Impairment and Williams Syndrome

  • Specific Language Impairment (SLI):
    • Affects grammatical development while leaving non-verbal IQ intact.
    • Likely has a genetic basis.
  • Williams Syndrome:
    • Causes general cognitive deficits, including in IQ, but leaves grammatical knowledge relatively intact.
  • Developmental Double Dissociation:
    • SLI and Williams Syndrome form a developmental double dissociation.
    • Suggests that general intelligence and linguistic competence are separate components of the mind.
    • Incompatible with empiricist views that link language directly to general intelligence.

Lecture 7: Large Language Models

  • Modern chatbots (e.g., ChatGPT) run on Large Language Models (LLMs).
  • Language models (LMs): Computer programs that predict the next word in a sequence based on training data.
  • Developed in Natural Language Processing (NLP), a subfield of artificial intelligence.
  • Limitations of LLMs:
    • Impressive achievements, but they don't yet offer a complete account of human language acquisition.
    • Require significantly more input than human children to train effectively.