Language Acquisition Quiz

Early Multi-Word Speech: Constructivist Approaches

• Multiword speech typically develops between 18 months and 2 years old.
• Two main theoretical approaches exist: nativist/generativist accounts and constructivist/usage-based accounts.
• This study focuses on constructivist accounts.

Overview

• Previous focus has been on learning to communicate using single words.
• This lecture explores how children combine words to form multi-word utterances.

Learning Objectives

• Describe the features of children's early multiword combinations.
• Outline two rule systems used to explain children's multiword speech.
• Explain the core ideas behind the constructivist (usage-based) account.
• Evaluate constructivist accounts of language acquisition using empirical evidence.

Lecture Structure

• Part A: Background (defining syntax, examining what needs explaining in multiword speech, and describing early multiword combinations and potential rules).
• Part B: Constructivist approaches – theory and initial evidence.
• Part C: Building an adult grammar.
• Overall summary and areas for evaluation.

Part A: Background

• 1. What is Syntax?: Syntax describes how words are combined in a language, enabling communication and creating an infinite number of potential sentences from a finite set of words. Examples are "who did what to whom?"
• 2. What needs explaining?: Language is a species-specific and universal human trait. Primates do not exhibit similar language acquisition capabilities, while children show advanced language skills by age 5.
• 3. What are early multiword combinations like?: Early combinations mostly consist of content words reflecting current situations (e.g., "Daddy shirt," "want drink"). Children also start to show some creative use of word order like adults (e.g., "more sing").
• 4. What kinds of rules could underpin early multiword utterances?: These rules can be lexical (word-based) or syntactic (grammatical).

1. What is Syntax?

• Syntax allows words to be combined in specific ways.
• Language allows understanding between speakers (e.g., who did what to whom).
• A finite set of words produces an infinite number of sentences.

2. What Needs Explaining?

• Language is species-specific. Very limited evidence of syntax acquisition in other primates through intensive training.
• Language is species-universal. Virtually all children acquire most of their grammar by age 5.

3. What are Early Word Combinations Like?

• Primarily content words describing actions/events in the immediate environment
• Show replication of adult word order
• Demonstrate some creative uses of word combinations

4. Lexical (Word-Based) Rules?

• Rules are item-specific, based on individual words or schemas (sets of words).
• Limited variety of utterances initially.
• Children need to generalize between schemas of words.

Syntactic (Grammatical) Rules?

• Rules are abstract, based on grammatical categories.
• Examples: Verb+Object, Subject+Verb
• Allow all possible utterances in adult language.

Interim Summary

• Children's early multi-word utterances are not random, but often based on previous experiences.
• Other species do not show comparable language capabilities.
• The study further explores how children combine words into sentences, the kind of knowledge needed, and the origins of this ability.

Part B: Constructivist Approaches: Theory & Initial Evidence

• 1. What is the Constructivist Approach (Usage-Based)? Grammar develops from communication, driven by infants' motivations to convey meaning. It is learned through general cognitive processes like communicative intention-reading, analogies, and distributional learning.
• 2. What kind of evidence would support the approach?: Evidence that children start with lexically-based linguistic representations. High-frequency words are learned first, followed by gradual generalisation across exemplars.
• 3. Evidence of Lexically-Based (Word-Based) Learning: The Verb Island Hypothesis: Children initially tie grammar to individual verbs. Generalization between verbs doesn't occur until late.

Experimental Evidence (Verb Island Hypothesis)

• Knowledge of grammar linked to specific verbs until 2½ years-old.
• Children struggle to apply grammatical rules to unfamiliar verbs.
• Children use more familiar verbs to show how actions are performed correctly.

Evidence: Limited (Lexical) Constructions

• Initial utterances are based on lexical items (words), not exclusive to verbs.
• Children use high-frequency words to create utterances.

Interim Summary (Children's Language Production)

• Good evidence that children's early utterances are more constrained than those of adults.
• Further investigation into how children 'join' various aspects of language development for sentence structure.

Part C: Building an Adult-Like Grammar

• 1. Structure Combining: Children build on previously said utterances to create new ones through simple changes like substitutions, additions, and deletions.
• 2. Semantic Analogy: Children initially learn multiple verbs and then generalize similarities across verbs. This process allows recognition of the actions related to various verbs.
• 3. Distributional Learning: Children learn cooccurrence patterns: how frequently certain words appear together in sentences (e.g., verb-ing/-ed/-s, plural/singular nouns).

1. Structure Combining (Lieven et al., 2003)

• Children build utterances based on preceding utterances.
• Studied a single child's utterances over several weeks.
• Examined recorded utterances and noted specific changes (substitutions, additions, deletions).
• Identified "target" and "source" utterances.

Method (for Structure Combining)

• Identify changes needed to transform a similar "source" utterance to a "target" utterance, noting operations needed.
  • Example operations: substitution, addition, deletion.

Results (of Structure Combining Study)

• High percentage of repetitions identified in the recorded data.
• A substantial number of original utterances discovered.

Conclusions (of structure combining)

• Many complex utterances are repetitions or small modifications of prior utterances.
• Modifications frequently involve simple word substitutions or single word additions/deletions.

2. Semantic Analogy

• Children learn various verbs.
• They notice similarities across verbs and create broader schemas.
• Dissimilar attributes of verbs are often ignored.

Evidence: Repeating Sequences (Matthews & Bannard, 2010)

• Tested 2 & 3-year-olds and assessed their ability to repeat 4-word sequences.
• Items related to each other in terms of meaning created fewer errors.

3. Distributional Learning

• Understanding how frequently words appear together in sentences and contexts.
• Identifying the use of Verb-ing/-ed/-s, Nuns-s'/s, plural/singular forms of nouns, and other grammatical patterns.

Experimental Evidence (Distributional Learning)

• Exposing children to multiple transitive sentences with familiar/unfamiliar verbs.
• Examining children's ability to apply noun/pronoun combinations in sentences.
• Testing children's ability to describe actions that involve novel verbs using grammatical structure components.
• Child's ability to use pronouns to increase the abstractness of the sentence.

Overall Summary

• Children begin combining words at 18-24 months of age.
• Constructivists suggest children figure out meaning by interpreting the intentions of others in predictable situations.
• Children use more limited rule systems (e.g., lexical rules) initially compared to adults and use general cognitive mechanisms.

Critical Evaluation

• Production studies are challenging for children due to memory demands involved in recalling words and planning sentences.
• The study raises questions about whether production studies underestimate children's abstraction of sentence structure capabilities.

Reading Guidance

• The studied material will be assessed through short-answer and multiple-choice questions.
• Understanding the material is crucial for success.
• Specific sections of Clark's book (2009/2016) are recommended for further reading concerning early language acquisition stages.