lecture 4 - morphological and syntactic development Flashcards

Question

Why should we care about theories?

Answer 1

Diagnosis Treatment Education Policies

Answer 2

images in notes

Answer 3

Nature - Internal forces * A priori knowledge * Rationalist position in philosophy * Innate, independent of experience * Final form exists prior to its emergence * Environment merely triggers pre- existing structures in the brain * Genetic determinism: “a gene for X” Nativism / Rationalism * Chomsky (1986) * Domain-specific (language-specific) * Universal grammar (UG) Emphasizes importance of: Genes Predicts: Language robust to cognitive deficits or impoverished input Nurture - External forces * A posteriori knowledge * Empiricist position in philosophy * Acquired by experience * Perceptual observation via the senses * Statistical regularities in the physical and social environments * Environment structures the brain Behaviorism / Empiricism * Skinner (1957) * Stimulus / response learning Emphasizes importance of: Environment Predicts: Language follows richness of input

Answer 4

But explaining the dynamics of interactions is no easy task….

Answer 5

* Learning the structure of language * Morphology * Syntax * Theoretical perspectives * Nature * Nurture * Interactionism!

Answer 6

- One of the most important issues in the study of language development is the extent to which our language abilities are innate. - There are 2 contrasting philosophical views on how humans obtain knowledge - The rationalists (plato and descartes) = certain fundamental ideas are innate so are present from birth - The empiricists ( such as locke and hume) rejected the doctrine of innate ideas and say all knowledge is derived from experience. - This is often called the nature and nurture debate - Chomsky’s work in general and his views on language acquisition are in the rationalist camp, and there are strong empiricist threads in Piaget. (Piaget argued that cognitive structures themselves are not innate, but can arise from innate dispositions.) - Behaviorists, who argued that language was entirely learned, are clearly empiricists. - The questions of which processes are innate and which must be in place for language to develop are of fundamental importance - Behaviour ultimately results from the interaction with nature and nurture - Work in connectionism has focused attention on the nature of nurture and the way in which learning systems change with experience (Elman et al., 1996). Language development is a complex process that involves the development of many skills and processes that may be important for syntactic development.

Answer 7

- The simplest theory of language development is that children learn language by imitating adult language. - Children clearly imitate some aspects of adult behavior, it is clear that imitation cannot by itself be a primary driving force of early language development, and particularly of syntactic development. - A cursory examination of the sentences produced by younger children shows that they do not often imitate adults. Children make types of mistakes that adults do not. - When children try to imitate what they hear, they are unable to do so unless they already have the appropriate grammatical construction. Imitation of adult speech (and that of other children) plays an important role in acquiring accent, in the manner of speech, and in the choice of particular vocabulary items. It might also be more important in older children.

Answer 8

- Chomsky (1965, 1968, 1986) argued that language acquisition must be guided by innate constraints, and that language is a special faculty not dependent on other cognitive or perceptual processes. It is acquired, he argued, at a time when the child is incapable of complex intellectual achievements, and therefore could not be dependent on intelligence, cognition, or experience. Because the language they hear is impoverished and degenerate, children cannot acquire a grammar by exposure to language alone. Assistance is provided by the innate structure called the language acquisition device (LAD) . - In Chomsky’s later work the LAD is replaced by the idea of universal grammar . This is a theory of the primitives and rules of inferences that enable the child to learn any natural grammar. In Chomsky’s terminology, it is the set of principles and parameter s that constrain language acquisition. - For Chomsky, language is not learned, but grows. Obviously languages vary, and children are faced with the task of acquiring the particular details of their language. For Chomsky (1981), this is the process of parameter setting. A parameter is a universal aspect of language that can take on one of a small number of positions, rather like a switch. The parameters are set by the child’s exposure to a particular language.

Answer 9

- Another way of looking at it is that the LAD does not prescribe details of particular languages, but rather sets boundaries on what acquired languages can look like; languages are not free to vary in every possible way, but are restricted. For example, no language yet discovered forms questions by inverting the order of words from the primary (declarative) form of the sentence. - The LAD can be thought of as a set of switches that constrain the possible shape of the grammars the child can acquire; exposure to a particular language sets these switches to a particular position. If exposure to the language does not cause these switches to go to a particular position, they stay in the neutral one. Parameters set the core features of languages. Thus this approach sees language acquisition as parameter setting. Let us look at a simple example. - In languages like Italian, it is possible to drop the pronoun of sentences. For example, it is possible just to say “parla” (speaks). In languages such as English and French, it is not grammatical just to say “speaks”; you must use the pronoun, and say “he speaks.” Whether or not you can drop the pronoun in a particular language is an example of a parameter; it is called the pro-drop parameter. - English and French are non-prodrop languages, whereas Italian and Arabic are pro-drop languages. But once the pro-drop parameter is specified, other aspects of the language fall into place. For example, in a pro-drop language such as Italian you can construct subjectless sentences such as “cade la notte” (“falls the night”); in nonpro-drop sentences, you cannot. Instead, you must use the standard word order with an explicit subject (“the rain falls”). - Pro-drop languages always permit subjectless sentences, so pro-drop is a generalization about languages (Cook & Newson, 2007).

Answer 10

According to Chomsky and others who view language acquisition as innately guided, learning a language involves setting internal grammatical "switches" or parameters. However, language development is slow and error-prone, raising questions about this model. Two explanations address the delay: Continuity hypothesis: All principles and parameters are present from birth but can't be fully used until cognitive capacities like categorizing words and processing longer sentences develop (Clahsen, 1992). Maturation hypothesis: Innate knowledge becomes accessible gradually through maturation (Felix, 1992). There is no consensus on which view is more accurate. Further challenges include the difficulty of identifying specific parameter settings across languages (Maratsos, 1998). For example, English-speaking children often omit pronouns in telegraphic speech, possibly due to incorrect parameter settings. Though this resembles Italian, the comparison fails since Italian verbs encode more subject information than English. Sign language acquisition poses another issue. Milestones in both signed and spoken language occur at similar times, and children make similar errors in both (Petitto, 1987, 1988; Newport & Meier, 1985). Yet, it’s unclear how sign language fits into a theory based on verbal parameters. Bilingualism adds complexity, as children must set parameters for two languages simultaneously (Messer, 2000). These difficulties led Chomsky to later downplay the role of abstract grammatical rules in language acquisition (Chomsky, 1995).

Answer 11

innate constraints on language must be general, applying across all languages—implying the existence of universal features. Chomsky argued that differences among languages are superficial, with underlying similarities forming a universal grammar. Pinker (1994) famously claimed that, to an outsider, humans appear to speak one language despite differing vocabularies. Though there are around 6,000 languages, they share a basic structure. Linguistic universals are features common across languages. Chomsky (1968) distinguished: Substantive universals: shared categories like nouns and verbs, even in children without linguistic input (e.g., deaf child “David” used different gestures for nouns and verbs; Goldin-Meadow et al., 1994). Formal universals: common rules governing how syntactic categories are manipulated. Greenberg (1963) found 45 universals related to word order in 30 languages. Notably, some word orders—like object-first—are rare, suggesting universal constraints. Primary word order (e.g., SVO in English) influences other aspects: SVO languages use prepositions and question-fronting, while SOV languages use postpositions and question-endings. Why do universals exist? Four main explanations: Innate grammar: Universals may stem from built-in grammatical structures, like parameter settings linking unrelated features. Innate cognition: Universals may reflect cognitive predispositions. For instance, infants recognize tight vs. loose fits (Hespos & Spelke, 2004), a distinction encoded in Korean but not English. Some concepts, like predicate-argument structure, may have neural roots (Hurford, 2003). Processing constraints: Certain structures are easier to process, shaping how languages evolve (Hawkins, 1990). Environmental influences: Universal features might arise from shared interactions with the physical world. However, Evans & Levinson (2009) challenged the idea of true universals, suggesting instead that languages vary in all possible ways. Developmentally, innate mechanisms emerge early, while learned grammar develops gradually. Yet Wexler (1998) proposed that some parameters (like word order) are set early, while certain universal properties emerge later through maturation. For example, by the two-word stage (around 18 months), children already grasp much of their language's inflectional structure.

Answer 12

- A study of a large number of same-sex twins found that vocabulary and grammatical abilities are correlated at the ages of 2 and 3, suggesting that the same genetic factors influence both abilities (Dionne, Dale, Boivin, & Plomin, 2003). Such results suggest that the innate basis of language is very general. - To some extent the debate is no longer simply about whether nature or nurture is more important, but about the precise mechanisms involved, and the extent to which general cognitive or biological constraints determine the course of language development. - Many people consider there is something unsatisfactory about specific innate principles. Having to resort to saying that something is innate is rather negative, because it is easy to fall back on a nativist explanation if it is not easy to see a non-nativist alternative. This is not always a fair criticism, but it is important to be explicit about which principles are innate and how they operate. - Innate principles are also difficult to prove. The best way of countering those researchers who see this as a negative approach would be to show where the principles come from and how they work: for example, by showing which genes control language development and how. - As Braine (1992) asked, exactly how do we get from genes laid down at conception to syntactic categories 2½ years later? We are a long way away from being able to answer this question. Nativist accounts tend not to give enough emphasis to the importance of the social precursors of language. It is possible that social factors can do a great deal of the work for which innate principles have been proposed. Researchers who are opposed to nativist theories argue that the learning environment is much richer than the nativists suppose: in particular, children are presented with feedback. Deacon (1997) argues that the structure of language itself facilitates learning it: Language has evolved so that it has become easy to learn.

Answer 13

An alternative to the idea of innate language knowledge is that children learn language from the input alone. Connectionist models support this by showing how complex linguistic behavior can emerge from exposure to language without needing built-in grammatical rules (Elman, 1999; Elman et al., 1996). These models emphasize the role of real linguistic input and general-purpose learning mechanisms (Gómez & Gerken, 2000). Infants appear to use distributional information—patterns of sounds and word sequences—to learn structure even without understanding meaning. Studies with artificial languages show children can learn linguistic rules from form alone, suggesting meaning need not come first. Elman (1993) showed that neural networks could learn embedded sentence structures—but only with gradual training, supporting Newport’s (1990) “less-is-more” hypothesis, where limited early cognitive resources may actually aid learning. Kersten & Earles (2001) found adults learned artificial languages more effectively when exposed to simpler segments first. However, introducing semantic content to the models (Rohde & Plaut, 1999) suggests that “starting small” may not always help with more natural input. Connectionist models suggest grammar can be learned from positive input alone, without explicit correction or innate grammatical knowledge. However, these models have not fully captured the complexity of natural languages. Whether language learning constraints must be innate (as per Gold’s theorem) or could arise from general brain functions or social environments remains debated (Elman et al., 1996). Despite this, both adults and children can extract syntactic patterns from statistical exposure. Saffran (2001, 2002) showed that even young children could learn structural dependencies in artificial languages. These skills also extend to non-linguistic and visual input, suggesting language learning relies on general cognitive mechanisms. In a key study, Marcus et al. (1999) found that 7-month-olds could learn abstract rules like the ABA pattern (“ga ti ga”) and detect violations (“wo fe fe”) from minimal exposure. This suggests infants can extract abstract structures from very limited input. However, there's ongoing debate about whether such behavior reflects true rule-learning or can be explained through simple statistical mechanisms (e.g., Christiansen & Curtin, 1999; Seidenberg & Elman, 1999; Marcus, 1999).

Answer 14

- A stage of single-word speech (called holophrastic speech) precedes a stage of two-word utterances. - Then early speech is telegraphic - grammatical morphemes may be omitted. - We can broadly distinguish between continuous and discontinuous theories - Continuous theories = children are believed to have knowledge of grammatical categories from the very earliest stages. Childs goal is to attach particular words to the correct grammatical categories and then use them with the appropriate syntactic rules. - Discontinuous theories = early multiword utterances are not governed by adult-like rules. Theoretical approaches also vary depending on the extent to which they emphasise the semantic richness of the early utterances.

Answer 15

Its a basic requirement of understanding and using language

Answer 16

Innate Knowledge Theory (Pinker, 1984, 1989) Children are born with innate knowledge of basic syntactic categories (e.g., nouns = objects, verbs = actions). They use semantic bootstrapping: Learn word meanings. Use these meanings + sentence structure to infer underlying grammar. Thematic roles help: Agent = doer (e.g., Vlad) Patient = receiver (e.g., Agnes) Linking rules map agent → subject, patient → object. ⚖️ Critiques of the Nativist View Input limitations: Children often don't hear clear mappings between agents/actions and words. Verbs, especially, can be hard to learn from context (Gillette et al., 1999). Acquisition patterns don't match predictions (Bowerman, 1990): “Easier” verbs (agent-action-patient) aren’t learned earlier than more complex ones. Children acquire different verb types at the same time. Braine (1988): Argued semantic input alone may be sufficient. Questioned early use of phrase-structure rules and need for innate syntax. Theoretical concerns: Innateness is a powerful claim—perhaps too powerful. Language learning is gradual and full of errors, suggesting more flexibility is needed. 📌 Summary Nativist accounts offer clear mechanisms (e.g., bootstrapping), but struggle to match actual learning data. Evidence suggests children may rely more on semantic input and distributional cues than on innate syntactic categories.

Answer 17

Core Idea Grammatical categories are initially based on semantic concepts, not syntax. Early language development is asyntactic—children begin by understanding meaning first (Gleitman, 1981; Macnamara, 1972). 🔍 Early Concept Formation Initial mappings: Nouns → objects Adjectives → attributes Verbs → actions ❗But: Many nouns and verbs refer to abstract ideas (e.g., "truth", "love", "think", "see"), challenging this simple mapping. Some researchers argue: Early semantic categories may underlie syntactic ones (McShane, 1991). Children may build the “noun” category from a semantic concept of objecthood (Gentner, 1982). 🧩 Theories Supporting the Semantic View 🧠 Semantic Assimilation Theory (Schlesinger, 1988) Early semantic roles (e.g., agent-action) gradually become syntactic categories. No need for innate syntactic knowledge. 🗣️ Macnamara's Lexicon-First Approach (1972) Children: Focus on individual content words first. Use context + word meaning to infer sentence meaning. Example: Seeing “Mommy drop ball” helps map “Mommy dropped the ball.” 🔁 Role of Sentence Schemas Default English word order: Subject – Verb – Object. Child-directed speech helps: Refers to the here-and-now. Uses simple syntax that aids understanding. ❓ Challenges to the Semantic View Verbs are tricky: Many don’t describe actions ("love", "want", "need"). Some adjectives describe states like verbs (“hungry”, “nice”). Yet children don’t make errors like saying “I hungries” → suggests syntax isn't purely semantically derived (McShane, 1991). 🧱 Grammar Emerges from Grammatical Properties 🧩 Maratsos (1982): Children identify syntactic categories via shared grammatical behavior. Examples: Nouns → appear first in declaratives. Verbs → take -ed or -ing suffixes. Once one category is learned, bootstrapping helps build others (e.g., adjectives come before nouns). 🔁 Bates & MacWhinney (1982): Abstract nouns are later absorbed into the noun category. Why? They behave like nouns (appear in same positions). Children use distributional cues to sort words into categories.

Answer 18

Core Idea Children can acquire syntactic categories using distributional cues with minimal semantic information. They treat language like a puzzle, solving it through pattern recognition. 🧩 Key Findings & Evidence Children as young as 2 acquire complex structures (e.g., gender inflections in Hebrew) despite weak semantic content (Levy, 1983, 1988). Children notice syntactic regularities before semantic ones. Children distinguish count vs. mass nouns using syntax: ✅ “a broomstick” vs. ❌ “a furniture” ✅ “much furniture” vs. ❌ “much broomstick” Children use syntactic (not semantic) cues to categorize (Gathercole, 1985; Gordon, 1985). 🔍 What is Distributional Analysis? The process of identifying syntactic categories based on the statistical distribution of words in sentences. Example: In “the X laughs,” X is very likely to be a noun. 💻 Modeling & Support Connectionist models support the theory: Show that categories can be learned statistically without innate grammar (Elman, 1990; Mintz, 2003). MOSAIC model simulates child language acquisition across multiple languages using only distributional analysis (Freudenthal et al., 2005, 2006). 📊 Real-World Application Children generalize from examples: ✔️ “Vlad eats fish” ✔️ “Vlad can fish” ❌ “Vlad can rabbits” Shows how syntactic acceptability can be predicted from distributional learning. ⚖️ Debates & Open Questions How much innate knowledge is required? New research suggests less is innate, since input is richer than previously thought. Children also use: Morphological cues (e.g., -s and -ed for verbs; -s only for nouns) Phonological stress (e.g., nouns stressed on first syllable, verbs on second)

Answer 19

- In summary, the relation between the development of syntax and the development of semantics is likely to be a complex one. - Early work emphasized the importance of semantic information in the acquisition of syntactic categories, but more recent work has shown how these categories can be acquired with little or no semantic information. Children probably learn syntactic categories through a distributional analysis of the language, and connectionist modeling has been very useful in understanding how this occurs. It is unlikely that innate principles are needed to learn syntactic categories.

Answer 20

- The failure of pure syntactic approaches to early development and the emerging emphasis on the semantic richness of early utterances led to an emphasis on semantic accounts of early grammars. - Aspects of Brown’s (1973) grammar were also derived from this: for instance, he observed that 75% of two-word utterances could be described in terms of only 11 semantic relations (see Box 4.6 for examples). - There is some appeal to the semantic approach in the way in which it de-emphasizes syntax and innate structures, and emphasizes mechanisms such as bootstrapping, but it has its problems. - First, there is a lack of agreement on which semantic categories are necessary. - Second, it is unclear whether children are conceptually able to make these distinctions. - Third, this approach does not give any account of the other 25% of Brown’s observed utterances. - Fourth, the order of acquisition and the emergence of rules differ across children. - Finally, Braine (1976) argued that this approach was too general: the evidence is best described by children learning rules about specific words rather than general semantic categories. - For example, when children learn the word “more,” is this a case of learning that the word “more” specifically combines with entities, or is it more generally the case that they understand that it represents the idea of “recurrence plus entities”? - If the latter is the case, then when children learn the word “more” they should be able to use other available recurrence terms (e.g., “another”) freely in similar ways; however, they do not. Hence the child appears to be learning specific instances rather than just semantic categories. - Braine gives the example of a child who learned to use “other” mostly only with nouns denoting food and clothing. He concluded that children use a combination of general and specific rules.

Answer 21

Learning syntax involves understanding which verbs take which arguments. For instance, “hits” is transitive (needs an object), “falls” is intransitive (no object), and some verbs can take both direct and indirect objects (e.g., “gives the ball to Agnes”). Development Pattern: Children show a U-shaped developmental trajectory: they initially use verbs correctly, then make overgeneralization errors (e.g., “Adam fall toy”), and eventually return to correct usage (Akhtar, 1999; Alishahi & Stevenson, 2005). This suggests early learning is based on specific examples, followed by over-applied generalizations, and finally refined understanding. Key Theories Explaining Verb Learning: Semantic Verb Class Hypothesis (Pinker, 1989): Children use verb meaning to determine structure. Verbs implying directional motion (e.g., “fall”) typically appear only in intransitive constructions. Overgeneralization happens before semantic distinctions are fully learned. Verb-Island Hypothesis (Tomasello, 1992, 2000): Early syntax is based on individual, frequently-used verbs (“verb islands”). Grammar emerges from analogies made between these islands. Evidence shows young children use verbs in only one construction type (Lieven et al., 1997), suggesting they learn item-by-item, not from abstract rules. Naigles’ Paradox (2002): Infants seem good at abstracting patterns (e.g., in comprehension tasks), while toddlers appear limited to specific items (in production tasks). Naigles suggests young children do abstract structures, but struggle to integrate them with meaning early on—form is easy, meaning is hard. Structural Priming (Savage et al., 2003): Children under 4 do not show priming effects like adults do, suggesting they lack abstract syntactic representations or may rely more on imitation. Entrenchment Hypothesis (Braine & Brooks, 1995): Repeated exposure to a verb in a specific structure reduces errors. High-frequency verbs are less likely to be misused than low-frequency ones (Theakston, 2004). Probabilistic/Distributional Models: Verb-argument structures are learned from frequent patterns and meanings. Children start with imitation and gradually form generalizations. Over time, they balance general patterns with specific usage (Alishahi & Stevenson, 2005). Syntax Abstraction: Early or Late? Late-Syntax Theories (Tomasello, Braine): Abstraction develops after 3 years of age. Early grammar is item-specific and shaped by lexical knowledge. Early-Syntax Theories (Pinker, Naigles): Abstraction occurs earlier. Children use innate or domain-general mechanisms to extract patterns. Conflicting findings are often due to methodology: Comprehension tasks (e.g., preferential-looking) show evidence of early abstraction. Production tasks suggest limited abstraction. Example: Preferential-looking studies show children under 3 understand who is doing what based on word order (Gertner et al., 2006), but act-out tasks do not. Connectionist models (Chang et al., 2006) bridge this gap: children may understand but struggle to produce correct structures, especially without external cues. Cross-Linguistic Insights Lidz et al. (2003) explored causative verbs in English vs. Kannada. In English, causativity aligns with transitivity; in Kannada, a morpheme indicates causativity. Findings: 3-year-olds in Kannada still relied on transitivity, despite the morpheme being a better cue. This supports universalist theories (innate syntactic biases), but emergentist accounts argue children simply follow the most salient cue in their input. Conclusion: Children learn verb-argument structure through a combination of specific examples, semantic insights, and gradually abstracted patterns. The debate continues between item-based vs. rule-based learning, and whether abstraction is innate, learned early, or constructed slowly through use and exposure.

Answer 22

- Can early syntactic development be both nonsyntactic and non-semantic? The identification of early syntactic categories might occur without much semantic help, and without being based on the acquisition of an explicit grammar. Instead, children seem to learn grammatical categories by distributional analysis. - Can this type of approach be extended to account for how children produce two-word and early multiword utterances? - Perhaps children’s early productions are much more limited than has frequently been thought (Messer, 2000). - Perhaps their early multiword utterances just statistically reflect the most common types of utterance they hear? - According to this view, children have a much less formal grammar than is commonly supposed. Evidence for this comes from the observation that early language use is much less flexible than it would be if children were using explicit grammatical rules (Pine & Lieven, 1997). - In general, the idea that there is a syntax module that drives language development is becoming less popular. It is clear that language development must be seen within the context of social development and the way language is used (Messer, 2000). - The shift is also mirrored in Chomsky’s more recent work (1995), where the importance of grammatical rules is much reduced. Perhaps there is no straightforward way of separating grammatical and lexical development; the two are intertwined (Bates & Goodman, 1997, 1999). - For example, grammatical development is related to vocabulary size: The best predictor of grammatical development at 28 months is vocabulary size at 20 months, suggesting that the two share something important (Bates & Goodman, 1999; Fenson et al., 1994). - Furthermore, there is no evidence for a dissociation between grammatical and vocabulary development in either early or late talkers: We cannot identify children with normal grammatical development but with very low or high vocabulary scores for their age. Neither is there any evidence of any clear dissociations between grammatical and lexical development in language in special circumstances (such as Williams syndrome and Down’s syndrome). Bates and Goodman (1999) concluded that there is little support for the idea of a separate module for grammar. In conclusion, recent work tends to downplay the role of an innate grammatical module and the attribution of adult-like grammatical competence to young children.

Answer 23

- Brown (1973) suggested that the mean length of utterance (MLU) is a useful way of charting the progress of syntactic development. This is the mean length of an utterance measured in morphemes averaged over many words. Brown divided early development into five stages based on MLU. - Naturally MLU increases as the child gets older; we find an even better correlation with age if single-word utterances are omitted from the analysis (Klee & Fitzgerald, 1985). - This approach is rather descriptive and there is little correlation between MLU and age after the age of 5. - Nevertheless, it is a convenient and much-used measure. The rule-based nature of linguistic development is clear from the work of Berko (1958). She argued that if children used rules, their use should be apparent even with words the children had not used before. They should be able to use appropriate word endings even for imaginary words. - In a famous study, Berko used nonsense words to name pictures of strange animals and people doing odd actions. For example, she would point to a drawing and say: “This is a wug. This is another one. Now there are two __” (see Figure 4.7). - The children would fill in the gap with the appropriate plural ending “wugs.” In fact, they could use rules to generate possessives (“the bik’s hat”), past tenses (“he ricked yesterday”), and number agreement in verbs (“he ricks every day”). The development of order of acquisition of grammatical morphemes is relatively constant across children (James & Khan, 1982). The earliest acquired is the present progressive (e.g., “kissing”), followed by spatial prepositions, plurals, possessives, articles, and the past tense in different forms.

Answer 24

- The way in which adults talk to children appears to have an effect as the child gets older: There are large individual differences in the ability of preschool children to form and understand syntactically complex sentences, and the quality of what children hear correlates highly with these differences (Huttenlocher, Vasilyeva, Cymerman, & Levine, 2002). - Children who hear complex structures master them earlier. Even here, it is difficult to be certain about what is causal. - The most important source of input for young children is their parents, so we cannot rule out genetic factors: Syntactic complexity in parent and child might reflect parent– child genetic similarity. However, the language of teachers also comes to have an effect: The syntactic abilities of children taught by teachers who use syntactically more complex speech develops faster than those taught by teachers who use simpler constructions (Huttenlocher et al., 2002). Hence language input does play a role.

Answer 25

- Languages differ in their syntactic complexity. For example, English is relatively constrained in its use of word order, whereas other languages (such as Russian) are more highly inflected and have freer word order. - Not surprisingly, these differences lead to differences in the detail of language development. What is perhaps surprising is the amount of uniformity in language development across languages. For example, stage 1 speech (covering the period with the first multiword utterances, up to MLU of 2.0) seems largely uniform across the world (Dale, 1976; Slobin, 1970). - There are of course some differences: Young Finnish children do not produce yes– no questions (Bowerman, 1973). This is because you cannot form questions by rising intonation in Finnish, so speakers must rely on an interrogative inflection. Some differences emerge in later development. - Plural marking is an extremely complex process in Arabic, but relatively simple in English. Hence plural marking is acquired early in English-speaking children, but is not entirely mastered until the teenage years for Arabic-speaking children (see McCarthy & Prince, 1990; Prasada & Pinker, 1993). - In complex inflectional languages such as Russian, development generally progresses from the most concrete (e.g., plurals) first to the most abstract later (e.g., gender usually has no systematic semantic basis; see Slobin, 1966b).

Answer 26

Jean Berko’s 1958 study is a foundational work in language acquisition research, exploring how children learn and apply morphological rules in English. Instead of relying solely on memorization, children show an ability to generalize and apply grammatical rules to new words, which demonstrates an understanding of language structure.

Answer 27

* The study investigates how children learn English morphology, particularly inflections like plurals, possessives, verb tenses, and adjective forms. * It aims to determine whether children have internalized grammatical rules or are simply repeating forms they have heard. The use of nonsense words (e.g., "wug") allows researchers to test whether children can apply rules to unfamiliar words rather than relying on memory.

Answer 28

Participants: * Children from preschool and first grade (ages 4–7). * Adult control group (college-educated speakers of English). * 56 children in the final sample, balanced for gender. Procedure: * Created nonsense words (e.g., "wug," "bik," "kazh") to ensure children were applying rules rather than recalling memorized forms. * Used picture cards to depict objects and actions. * Presented structured prompts where children had to provide the correct morphological form (e.g., plural, past tense). Collected quantitative data on correct vs. incorrect responses and analyzed errors for patterns.

Answer 29

A. Plural Formation (Regular and Irregular Plurals) * Most children correctly applied the standard plural rule (-s, -z, -ɪz). ○ Example: "This is a wug. Now there is another one. There are two __?" → "wugs" (correct). * Children struggled with the more complex allomorph (-ɪz), used after sibilant sounds. ○ Example: "This is a niz. Now there is another one. There are two __?" → Many children incorrectly said "niz" instead of "nizzes." * Errors reveal a simplification strategy: ○ Children tended to add -s or -z but had trouble recognizing when -ɪz should be used (as in "glasses" or "buses"). B. Past Tense Formation * Regular past tense (-ed) was applied correctly in most cases. ○ Example: "This is a man who ricks. Yesterday, he __?" → "ricked" (correct). * Children did not generalize irregular past tense patterns. ○ Example: "This is a man who rings a bell. Yesterday, he __?" ○ Instead of saying "rang," children incorrectly said "ringed." * Children had difficulty with phonologically complex past tense formations (-ɪd). ○ Example: "Yesterday, he __?" (for "melt") → "melted" (correct), but errors appeared for less familiar forms. * Overall: Children first learn the regular -ed rule, and irregular forms develop later with more exposure. C. Possessive Formation (Singular and Plural) * Children correctly formed singular possessives (-'s) more often than plural possessives (-'). ○ Example: "This is a wug. This is the wug’s hat." (correct). ○ Example: "These are two wugs. These are the __ hats?" → Many children struggled to say "wugs’ hats." * Plural possessive errors: ○ Children often omitted the possessive marker for plural nouns, treating "wugs" as both singular and plural. * Possessives were mastered earlier than complex plural forms, suggesting they are easier to learn. D. Third-Person Singular Verb Forms * Children applied the third-person singular (-s) rule inconsistently. ○ Example: "This man knows how to nazz. Every day, he __?" → Many failed to say "nazzes." * Performance on third-person singular was worse than on noun plurals. ○ Suggests that verb morphology may take longer to acquire than noun morphology. E. Comparative and Superlative Adjectives (-er, -est) * Children had trouble forming comparatives and superlatives. ○ Example: "This dog is quirky. This dog is __?" (quirkier) → Many children repeated "quirky" instead. ○ Few children successfully generated "quirkier" or "quirkiest." * Implication: Adjective inflection may develop later than noun/verb morphology. F. Derivation and Compounding * Children preferred compounding over derivation. ○ Example: "A man who zibbs for a living is a __?" → Instead of saying "zibber," children often said "zibbing man." ○ Example: "A small wug is a __?" → Instead of saying "wuglet," children said "baby wug." * Compounding was used more frequently than suffix-based word creation. ○ Suggests that derivational morphology (adding suffixes like -er, -y) develops later. G. Understanding of Compound Words * Children showed partial awareness of word structure but often assigned "private meanings." ○ Example: "Why is a blackboard called a blackboard?" § Some said, "Because you write on it" (functional answer). § Others said, "Because it is black" (focus on one element). * Some children created incorrect but logical meanings. ○ Example: "Friday is called Friday because we eat fried fish." Example: "Breakfast is called breakfast because you eat it fast."

Answer 30

A. Children Learn Rules, Not Just Words * Findings confirm that children don’t simply memorize words—they internalize patterns. * The ability to apply rules to nonsense words shows that children acquire abstract linguistic knowledge. B. Morphological Development Follows a Sequence * Simpler, more regular forms (e.g., plural -s, past tense -ed) are acquired first. * Complex or irregular forms (e.g., -ɪz plurals, strong past tense like "rang") are learned later. * Derivational morphology (creating new words using suffixes) develops after compounding. C. Learning is Based on Frequency and Exposure * Children apply common rules first (e.g., regular plurals and past tense) because they hear them more often. * Less frequent forms, like irregular verbs, take longer to master. D. Morphological Learning is Not Solely Imitation * Children’s errors are systematic, not random, indicating they are actively constructing grammatical rules. Overgeneralization (e.g., "ringed" instead of "rang") shows that children apply rules even when incorrect.

Answer 31

* Jean Berko’s study was groundbreaking in demonstrating that children actively construct and generalize morphological rules rather than just mimicking adult speech. * The Wug Test remains one of the most famous experiments in linguistics, supporting theories of rule-based language acquisition. * Future research expanded on Berko’s work by exploring bilingualism, cognitive development, and neurolinguistics in morphology learning. This study laid the foundation for modern research on child language acquisition, showing that morphology learning follows predictable developmental patterns.