Notes: One Year to One Year Six Months: Prerepresentational Phonology
1st 50 word period
This period presystematic, i.e. child's productions lack sound contrasts and/or consistent rules or
processes, and has great phonetic variability.
Children appear to produce whole words rather than dividing words into their component segments.
They are using sounds phonetically, but not necessarily contrastively, e.g. we may hear [ba] and [da]
and assume that the child has mastered the phonemic contrast /b/ and /d/, but if we listen further we
may see that the child is using the two sounds interchangeably when referring to the same object, e.g.
both [ba] and [da] for "daddy".
1st words characterized by simple syllable structures, typically CV (consonant-vowel) and VC, and
some CVC syllables; some show a great deal of syllable reduplication in CVCV productions.
Phonetic repertoire consists primarily of the manners of production of stops, nasals, glides; place of
articulation is commonly labials and alveolars, which are more common than palatals and velars.
Preferred vowels seem to be concentrated on the basic vowel triangle of [i] (i.e. "eee"), [a] (i.e. "ah"),
and [u] (i.e. "ooo").
Also, begin to see selectivity or avoidance patterns: attempts at word productions based upon the
word's phonological characteristics. Patterns of selection / avoidance are regulated by particular
phonemes, consonant clusters, or the length of target words. This is referred to as "lexical selection,"
and considered a problem solving activity for managing the challenge of the phonology.
Systematic Development: Representational Phonology
Phonemic Development (1 year 6 months to 4 years):
Clinically, our detection and treatment planning for speech sound disorders is most often based on the
order of acquisition of phonemes and age norms in development. Researchers in phonological
development develop tables of normative data regarding the ages at which children develop aspects
of the phonology by testing the consonant phonemes (and sometimes the vowels) and reporting the
percentage of children that produce the adult target sounds accurately at a specific age.
Studies demonstrate the same pattern over and over again in that stops, nasals and glides are
mastered earlier, and are followed by liquids, fricatives and affricates.
However, differences in age norms are observed when comparing studies. This is partly due to the
criteria used to define "acquired."
o Some studies (Templin, 1957; Prather, et. al., 1975) considered a sound as acquired when it is
produced correctly by 75% of the subjects at any age level. On the other hand, Poole (1934)
required 100% of subjects as the criterion.
o Another reason for the differences in the reported data is the word positions considered by the
investigators: Templin (1957) required accuracy of production in all 3 word positions (initial word
position, medial word position, final word position) whereas Prather looked at only initial and final
o Yet another reason for the differences in reported data comes from the fact that different
researchers required different degrees of phonetic accuracy of the production. /s/ a good example:
if strict criterion for phonetic accuracy is applied, /s/ would be considered a late acquisition, as
children take some time for a fully accurate production of this phoneme. On the other hand,
children reveal their awareness of this phoneme by approximating its production rather early
despite their difficulties in its phonetic implementation. Thus, it might be considered an early
acquisition if we disregard the imperfections in the productions. It has been suggested that
phonemic acquisition is ahead of phonetic acquisition for fricatives, i.e. children are aware of the
meaning contrasts before they produce them accurately.
o Another variable contributing to the lack of agreement among normative studies: types of
utterances used in testing. Some studies sampled speech sounds in the production of isolated
words, others studies in connected speech; length of words differed; stress patterns differed, word
familiarity differed, number of words tested for each word position differed, the effects of sounds in
words such as assimilation an unaccounted for variable; and, the conditions under which data was
collected differed—all are potentially influential factors in the resulting generalizations.
Consonant Clusters: a late development, but some develop earlier than others.
o /s+nasal/, /s+stop, and /stop+liquid or glide/ clusters developed earlier in the initial position of
o The final position favors the combinations of /nasal or liquid + voiceless stop/ and /stop + stop/ or
/stop + fricative/.
o 2 member clusters acquired earlier than 3 member (e.g. /st-/ before /str-/.
Vowels are generally mastered by 3-4 years; r-colored vowels (e.g. "er" or "ir" as in "bird) acquired
Years 4-7: Phonetic Inventory Completion
certain contrasts such as fricatives and affricates stabilized
by end of this period, child capable of producing all of the English sounds
still has difficulty with longer words such as "thermometer," "vegetable."
morphophonemic development also begins in some structures, e.g. /s/ and /z/ for the plural morpheme
The final stage of phonological development relates to the acquisition of a system of rules for the
combination of morphemes, e.g. vowel alternations such as sane-sanity, succeed-success, decide-
decision, tone-tonic. Also acquire stress alterations important for distinguishing compounds, e.g. `im-
port vs. im-`port, etc.
A "natural" theory of phonology in that it presents the sound system of the language as a natural reflection
of the needs, capacities, and world of its users rather than as a merely conventional institution.
Based on the notion that some phonemes are more natural to the organism than others, therefore
they will appear earlier in development and in more languages across the world.
Phonological Processes: operations that affect sound change, both within individual language users and
within the history of a particular language. Affect classes of sounds. Phonological processes are used by
children to "simplify" the sound system until they develop the capacity to better produce the adult model of
the system. For example, a child may substitute /w/ for /r/ for some time during development, as in the
production of "wabbit" for "rabbit." Later, the child replaces the /w/ with the target /r/. It is said that during
the production of w/r, he is engaged in the use of the phonological process of "gliding," i.e. substituting a
glide for the rhotic /r/.
Naturalness: more natural sounds appear before other sounds in language development; are heard
more often universally across languages.
Markedness: more marked sounds occur later in language development, and appear less often in
Assimilatory processes: processes in which a sound changes to become more similar to another
Some Theories of Phonological Development
Mower, 1952 & 1960; Olmstead, 1966 & 1971.
Emphasizes the role of contingent reinforcement in phonological acquisition.
Natural Phonology theory
Stampe, 1969 & 1973
Proposes that children do not actually acquire or develop a phonological system; instead, they learn to
suppress or constrain processes that do not occur in the language.
Macken & Ferguson, 1983; Menn, 1983
Point out that other theories don't account for individual variability in children acquiring the same
Children are active problem solvers
Locke, 1983 & 1990
3 basic premises:
o Prelinguistic vocalizations of infants are universal
o Phonological patterns of early meaningful speech closely resemble those of late babbling
o Substitutions are usually frequently occuring sounds in babbling repertoire
Stages in Biological theory
1. a prelinguistic stage: includes proto-words
2. child attempts to produce conventional words
3. increased similarity to adult speech with increased phonological complexity
Developmental speech sound errors made as a result of "weak connections" between target sounds
and needed features for their production.
Notes: Lexical Development
Lexicon: our mental dictionary in which knowledge of words and their meanings is kept. Average 1st
grader maintains about 14,000 lexical items; average English-speaking college student around
What is a word?
o A word is a sign that signifies a referent, but the referent is not the meaning of the word. If you say
to a child, ―Look at the kitty,‖ the referent—the actual cat—is not the meaning of kitty; if the cat ran
away, the word would still have meaning because meaning is an act of cognition.
o Cats can be called a number of words, depending upon the language spoken. There is nothing
intrinsic to cats that makes one or another name more appropriate—the relationship between the
names and the thing is thus arbitrary; it is by social convention in a particular language that
speakers agree to call the animal by a particular word. This arbitrary relationship between the
referent (the cat) and the sign for it (the word cat) is symbolic. Nonverbal signs can also be
symbolic, e.g. we could agree that a blue light means stop and change all the red lights to blue.
o In a few words, the relation between word and referent is not so arbitrary, e.g. thud, cuckoo clock.
Note that many of children‘s earliest words or protowords have a less-than-arbitrary relation to
their referents, e.g. clocks are called tick-tock, and painful bumps are called owies.
One of the child‘s primary tasks in semantic development is to acquire categorical concepts (e.g. to
learn that the word dog refers to a whole class of animals). There are three theories regarding how
children acquire categorical concepts.
o The semantic feature view: children learn a set of distinguishing features for each categorical
concept. When a child learns a new word, it is in the context of a specific situation, e.g. the word
―dog‖ may at first be understood to apply only to the child‘s own dog, and only later comes the
understanding that other creatures may also be called ―dog‖ so long as they share the same
critical features that uniquely define the category: alive, warm-blooded, have 4 legs, bark, covered
with hair. Overextensions occur when the child infers a word belongs to a category from a partial
match of features, e.g. a toddler may refer to a moose as a ―doggie,‖ because both animals have
hair and 4 legs. According to feature theory, the child in this case does not yet know that antlers
disqualify an animal from membership in the dog category.
o It may be that children acquire probabilistic concepts: most birds have feathers and beaks, fly,
chirp, etc., but not all do; e.g. kiwi is a nonflying bird, as is a penguin. Conclusions regarding
reference are based upon probability.
Some examples of probabilistic concepts have more of the qualities of the concept than others,
e.g. a robin has more typical ―bird‖ characteristics than does a penguin—therefore, people see
robins as better examples of birds, and they also classify them faster when asked if a robin is a
bird. These typical examples of the category, or prototypes, are more accessible in memory in
adult subjects. According to prototype theory, children acquire these core concepts when they
acquire meaning and only later come to recognize members of that category that are distant
from the prototypes.
The Course of Early Lexical Development
o Produced between 10 & 15 months of age. Often hard to distinguish from protowords; even though
protowords appear to have consistent meaning for children, the sounds of them don‘t relate to real
words in the language. In contrast, true first words are approximations of words in the target
language, even if rough.
o Many first words are context-bound, e.g. text example of Allison Bloom producing the word ―car‖
only when looking out her apartment window to the street, but not when seeing a picture of one or
when up close to one; the child presumed to have identified one particular event in the context of
which it was appropriate for her to produce the word ―car.‖ Other examples: ―more‖ might be used
only as a request and not to comment on recurrence, and ―no‖ might be used only as a refusal.
This is situation-specific or function-specific word use, very different from adult representations.
o Some researchers argue that context-bound words appear first and that truly referential words
must await some cognitive development in the child. Other suggest otherwise.
o First words can be referential. Harris and colleagues found that the largest category of words
produced was context-bound (22 of the first 40 words). The next was contextually flexible nominals
(names of things) accounting for 14/40; there were also 4 contextually flexible words that were not
nominals, e.g. more, yes, no.
o Why are some context-bound and some referential? Children seem not to make use of the full
range of their linguistic experience.
o Context-bound words become decontextualized.
From first words to 50 words
o Reach 50 around 18 months, range 15-24 months. Nelson classified these into 6 categories:
Specific nominals, e.g. Mommy, Daddy, Rover
General nominals, e.g. dog, ball, milk, he, this
Action words, e.g. go, up, look
Modifiers, e.g. big, all gone, outside, mine
Personal social words, e.g. no, want, please
Grammatical function words, e.g. what, is, for
o both specific and general nominals most common
o Most frequent nouns include names of people, food, body parts; also, words for clothing, animals,
household items involved in child‘s daily routines. Not universal, though; e.g., verbs are more
prevalent in Korean, and other East Asian cultures the nouns, which are more common in Western
o In noun-heavy early vocabularies, it was thought that object labels were easier to learn because
verbs express relationships among things, requiring more sophisticated cognitive function.
However, according to cross-linguistic and cross-cultural studies, in the Asian languages verbs
often come at the ends of sentences, so more salient; also, these languages allow noun dropping,
making verbs relatively more frequent in the language. Also, in American culture, caregivers spend
a lot of time labeling things for their babies. Also, nonlinguistic context provides more information
about noun meaning than about verb meaning.
At the 50 words stage and beyond
o lexical development shifts into a different gear at around 50 words. Rate at which they learn new
words increases from about 8-11 per month to now 22-37 per month. Word spurt.
o In English, it is a spurt primarily in the acquisition of object labels. Likely that this occurs because:
1. reaching 50 words gives the child a basis for figuring out principles of how a lexicon works, 2.
phonological abilities influence spurt, 3. ongoing cognitive development.
o Evidence for word comprehension begins at age 5 months; respond first to their own name; 8
months understand a few phrases; at 16 months, comprehension vocabulary of 90 to over 300
Individual differences in lexical development
o Input important: when words are taught as labels and given explicit definitions, they are more likely
to be used referentially from the beginning than those the child picks up from context; if a word
used is almost exclusively in a single context, the child has no basis for figuring out the more
o Child‘s approach to the language acquisition task constitutes individual differences as well: some
are more analytic than others; risk-takers; how social they are.
o Nelson described children with more object labels in their vocabularies as referential, whereas
others have more personal-social words: they are expressive. Most children fall in-between. May
depend on how mothers teach. Also, inherent differences in children: some more interested in
objects, others more interested in social interaction. Or may be an analytical vs. holistic distinction
(analytic being more referential).
o Regarding rate of lexical development:
Phonological memory as a cognitive skill
Gender: girls faster than boys (small difference)
Differences in temperament: outgoing child may elicit more input from others
First-born may have slight advantage; mothers talk more
More educated parents talk more to children
When rich in structural cues
How are new words learned?
o Speech segmentation issue: how does the child isolate ―cup‖ from ―Thisisacup‖? Stress and
rhythm good markers of word boundaries. In English, a stressed syllable is likely to signal the
beginning of a new word; so, if child pays attention to the contrast between stressed and
unstressed syllables in the speech they hear, they will be able to identify word boundaries. Child-
directed speech important.
o The mapping problem: how does child develop a concept of word meaning? What about the cup
makes cupness? There are an infinitie number of possibilities. Fast-mapping: between a new word
they hear and a likely candidate meaning. Develop hypothesis. Use assumptions that constrain the
number of possibilities.
The whole-object assumption: words refer to whole objects. E.g., ―hot‖ refers to stove.
The taxonomic assumption: words refer to things that are of the same kind. A taxonomy is a
system of classifying things into categories. When the child hears ―dog‖ in the presence of a
dog (and assumes that the whole object is being referred to), then the taxonomic assumption
leads the child to think that dog will also refer to other dogs but not to things that are
thematically related to dogs, such as collars, leashes. Children know that words work this way,
not e.g. toys, which they prefer to sort associatively, e.g. car gets grouped with person, not
The mutual exclusivity assumption: different words refer to different kinds of things. So,
members of the category ―dog‖ don‘t overlap with members of the category ―cow‖.
o Pragmatic principles as a source of support
Principle of conventionality: the meaning of a word is determined by convention
Principle of contrast: different words have different meanings (different from mutual exclusivity
because it allows an object to have more than one label with different meanings, e.g. dog and
o Input as a source of support
Speech directed toward young children is in the here-and-now. So, very little ambiguity. Adults
follow the children‘s attentional focus and encode what they‘re attending to.
o Sociopragmatic cues to word meaning
Children match words with speakers‘ intended meanings. The child may follow the speaker‘s
o General learning processes as the basis for word learning
General processes of attention and learning can explain what constraints and sociopragmatics
intend to explain. Attend to similar characteristics (shape), salient things, and attach words to
o Syntax as a clue to word meaning
Once children have acquired enough grammar to identify nouns, verbs and adjectives, they can
use the grammar as a clue to meaning, e.g. –ing likely a verb, prior article indicating noun, verb
between subject and object, etc. This is the syntactic bootstrapping hypothesis: the assumption
that knowledge of language structure is generally useful for learning new verbs.
The Relation of Words to Concepts
Have discussed how words become related to concepts the child may already possess. But, how do
concepts themselves develop?
Meanings of words come from the concepts they encode.
We most commonly view the task of the child as mapping the words heard onto a preexisting
conceptual structure; so, the kinds of word meanings children attend to are a function of the mental
categories children have.
Children possess ontological categories: mental categories from which they organize the world and
word meanings, e.g. substance vs. object categories to guide their inductions of word meaning;
cognitive categories as a basis for inferring word meanings.
But, lexical organization is different from conceptual organization, i.e. words do not always map onto
concepts one to one. Words mark some but not all conceptually available distinctions, e.g. Spanish
has no distinction in words for fingers and toes; some languages have fewer than 6 color terms.
So, when children are learning the lexicons of their languages, they must determine which cognitive
distinctions are marked in their language and which are not.
o The level of organizing the world that mediates between cognitive organization and language is
called semantic organization.
o Early, fast mapping occurs: only a partial mapping of the new word onto a conceptual domain. May
know that a thing has its own name even though they don‘t know the name. This is very early
o In Spanish, motion verbs generally encode path of motion, whereas in English they encode
manner (e.g. leaving vs. running)
o In English, we describe path of motion with a different word, e.g. I fell down, whereas in Korean the
direction of motion is part of the meaning of the verb (like ascend and descend): path is lexicalized
along with path; in English, path is lexicalized independent of path.
o These are examples that kids must figure out (and they do, very early) how words divide meaning
into lexical components.
Do words influence conceptual development?
o E.g. Spanish does not have different words for fingers and toes. Do Spanish speakers think of
fingers and toes as more similar than English speakers do? Perhaps cognitive organization doesn‘t
always come first; language might influence cognition.
o This is the linguistic relativity hypothesis, or the Worfian hypothesis (Benjamin Lee Worf): the
way our language ―carves up‖ the world influences how we think about the world. Studies
regarding the thought processes of people who speak different languages resulted. Some studies
have shown that characteristics of the language children are acquiring exert an influence on their
o Some languages may make it easier to learn verbs than others, e.g. Korean and Japanese, since
verbs more frequent and salient. So, these children, according to the hypothesis, should acquire
concepts encoded by verbs at a younger age.
o Brown demonstrated that parents responding to the child‘s ―What‘s this?‖ with ―maple‖, ―dogwood‖
or ―oak‖ rather than ―tree‖ may develop more differentiated mental categories of trees.
o Also, labels may be invitations to form categories.
Probably, words and concepts develop together
o Concepts don‘t develop and wait for words; nor do words prompt the development of the concepts.
o Close relationship between when children develop a correspondence between the appearance of
words that encode disappearance, e.g. ―gone‖, and the age at which children understand that
objects are permanent.
o Gopnik and Meltzoff proposed that children acquire words that encode concepts they have just
developed or are in the process of developing.
o Perhaps even the beginning of word learning itself coincides with the infant‘s mental individuation
of the different sorts of objects they encounter. E.g. study in text that suggested that 10 month olds
perceive things as ―objects‖ but not as separate objects, whereas 12 month olds did (cup and book
behind screen example)
Later Lexical Development
o After early childhood, 1. growth in vocabulary size, 2. growth in knowledge of word formation
processes, 3. increasing ability and importance of being able to learn new words from context.
o Regarding growth in vocabulary size:
Grade 1: over 10, 000 words
Grade 3: nearly 20,000 words
Grade 5: 40,000 words
Accounting for this growth: 1. root words; 2. inflected words: root plus a grammatical inflection,
e.g. boys, soaking; 3. derived words: roots plus an affix, e.g. sadness, preacher; 4. compounds,
e.g. payday, milk cow; 5. idioms, e.g. carrying on as meaning ―misbehaving.‖ Big increase from
3rd to 5th grade in derived words (recognition). This does not reflect an increase in number of
words known; rather, it reflects an increase in the child‘s ability to figure out words they have
never heard before, using their knowledge of root and affix meanings. So, the most important
development here is in morphological knowledge.
E.g. Berko‘s ―wug‖. Plural is ―wugs‖; past tense of ―rik‖ is ―riked‖ (study 4-7 years old). So, she
tested inflectional morphology. If you asked a child of this age what you call a tiny wug (wuggie,
wuglet, wugling), more difficulty, since this involves derivational morphology. Inflectional
morphemes add to a word; derivational morphemes make a new word: knowing how to build
new words by adding morphemes to other words, or knowing how to interpret newly
encountered words by recognizing their component parts. Creating ―wughouse‖, compounding
word-formation process, requires knowledge of how to combine 2 existing words.
Children possess these abilities early, but do not possess full control. Knowledge of
derivational morphology appears later than compounding.
Children know that inflectional morphemes are added to words after derivational morphemes
but not the other way around. E.g. what do you call someone who eats rats? Rat-eater, not
rats-eater, as you can‘t have a plural inside a compound. Can make mice-eater, though.
Learning words from context
o Incidental learning: 5-year-olds benefit from incidental exposure, not 3-year-olds. Better at
reference; 3-year-olds need a present reference.
o Learning words from context involves identifying all the relevant cues to converge on a definition of
an unknown word, and drawing on background knowledge.
o Could also use explicit vocabulary instruction.
Syntactic and Morphological Development
After first words develop, vocabulary begins to grow quite rapidly, and children begin to use their first
words in a variety of contexts, but their messages are limited to speaking one word at a time.
Usually in the latter half of the second year, they reach next important milestone: they begin to put
words together to form the first "sentences."
This is a critical turning point; even the simplest two word utterances show evidence of syntax: the
child combines words to create sentences following certain rules rather than in random fashion.
Syntax is the component of grammar that governs the ordering of words in sentences.
Takes place with no explicit instruction; parents more focused on what the child is saying rather than
how he says it.
The most influential linguistic frameworks characterizing the rules that underlie the well-formed
sentences of adult language users - the natural end point of the acquisition process - developed by
Noam Chomsky, called Universal Grammar.
o Began developing it in 1957
o Several revisions since
o Current version known as government and binding theory (1981)
o Any theory of grammar must be universal, and must meet the goal of learnability, i.e. that children
worldwide acquire the grammar of their language within a few short years and without explicit
training or correction.
o A theory of how we represent language as a set of principles in our mind. Chomsky believes our
mental representation of grammar is autonomous of other cognitive systems; so, the principles and
rules of grammar are highly specialized.
Central to this theory is that there are several components of the grammar that are linked at different
levels of representation.
o Of key interest are 2 levels: 1. deep structure, capturing the underlying relationships between
subject and object in a sentence, and 2. surface structure, capturing the surface linear
arrangements of words in a sentence.
o Example: John is eager to please; John is easy to please.
Both sentences have virtually the same surface structures: noun-verb-adjective-infinitive verb.
But they mean quite different things. The subject of the verb "to please" is John in the first
sentence, but someone else in the second. This difference in the underlying grammatical
relationships of subject, predicate, and so forth would be captured by very different deep
So, how do children come to grasp the underlying grammatical relations of sentences they hear (deep
structures) when they are only presented with surface structures?
Each structure level has several components.
o Surface structure has 2 parts: phonetic form (the actual sound structure of the language) and
logical form (capturing the meaning of sentences, connecting grammar to other aspects of
o Deep structure also has 2 parts: phrase structure rules, capturing the basic subject/predicate
structure of a sentence; and the lexicon, specifying morphological and syntactic features for each
lexical item in a sentence. Together, the lexicon and phrase structure rules generate the deep
structure of a sentence.
Phrase structures are often represented in tree-diagrams. They specify the underlying relationships of
parts or phrases of the sentence, e.g. noun phrases, verb phrases, adjectival phrases, etc.
o includes some additional syntactic elements such as complementizers (that, what), which
introduces each sentence, and an inflectional category, which holds the auxiliary verb (do, will,
may, etc.) and carries information about tense.
o Thus, the basic structure of the sentence is organized in the deep structure by the phrase structure
o And within these phrases, 2 types of categories:
Lexical category: headed by lexical forms such as nouns or verbs
Functional category: a grammatical category, such as inflectional or complementizer.
the lexicon provides the specific words that get inserted at the end of the phrase structure trees.
Contains information for each item about its syntactic category (e.g. is it a noun, verb, adjective, etc.),
and contains information about what kinds of sentence structures the item requires, especially
important for verbs (e.g. "run" only requires a subject; "see requires both a subject and an object; "put"
requires a subject and object, and also requires a location to be specified.
The deep structure is connected to the surface structure by a rule that reorders the elements of the
phrase structure into the linear arrangement of the surface form. This is called a transformational rule.
This moving around rule important in English, e.g. for making questions.
Grammar also has to have constraints regarding which elements can be moved and which can't be,
as well as where they can be moved to.
Measuring Syntactic Growth
Age itself is not a good predictor of language, especially syntactic, development since children
develop at vastly different rates.
The length of a child's utterances seems to be an excellent indicator of syntactic development; each
new element of syntactic knowledge adds length to a child's utterance.
Roger Brown (1973) introduced the major measure of syntactic development: the mean length of
utterance (MLU), based on the average length of a child's sentences scored on language transcripts.
Length determined by the number of meaningful units in the sentence: morphemes, rather than words.
Morphemes include simple content words, e.g. cat, play, do, red; function words e.g. no, the, you, this;
and affixes or grammatical inflections e.g. un-, -s, -ed, etc.
The addition of each morpheme (or minimal unit carrying meaning) reflects the acquisition of new
linguistic knowledge. So, children who have similar MLUs are at the same level of linguistic maturity,
and their language is considered to be at the same level of complexity.
In longitudinal studies, MLUs calculated at successive points in time gradually increase. Grows at
different rates in different children.
According to Miller and Chapman's norms (1981), Adam and Sarah about normal for their age.
Using MLU, Brown subdivided the major period of syntactic growth into 5 stages, beginning with stage
I when the MLU is between 1.0 and 2.0. Successive stages are marked by increments of .5; thus,
stage II goes from 2.0-2.5, and so forth up to stage V. Beyond an MLU of 4.0 some of the
assumptions on which the measure is based are no longer valid, and longer sentences do not simply
reflect what the child knows about language; so MLU loses value as an index of language
development after this stage.
Considerable problems are encountered in measuring the MLU in foreign languages, especially highly
inflected and synthetic languages such as German, Russian or Hebrew. Becomes difficult to decide
what functions as a morpheme in the child's speech, and it is easy to obtain inflated numbers.
The units that are combined in language are smaller than words themselves; they are morphemes. 2
o Bound morphemes: morphemes that don't stand by themselves
o Free morphemes: units that can stand alone
o Inflectional morphemes add grammatical information to words, but don't change meanings of
o Derivational morphemes form new words, e.g. the –er in dancer, the –ish in smallish. Word type
o Other languages have much more complex morphology.
descriptive rules: linguists take whatever people do as "correct" and try to describe the patterns in it;
way of studying language acquisition. Prescriptive rules are grammar rules teachers teach us. So,
how do children learn these descriptive rules?
Children's first word combinations are simple, active, declarative structures; missing function words
and bound morphemes. Negative sentences and questions come later. Lastly, multiclause sentences
appear. Process relatively complete around age 4.
Syntax & Morphology II
2 types: yes/no and wh- (including "how")
o 1st yes/no marked by intonation only, e.g. "More juice?" for "May I have more juice?"
o wh- questions really a statement with a wh- word attached to the beginning, e.g. "What that is?"
(i.e. no reversal)
o Next, add auxiliaries to beginning of yes/no questions, e.g. "Will it fit in there?" wh- questions do
not yet invert the subject and auxiliary, e.g. "What a doctor can do?"
Development of complex sentences
Late. There are different types.
First complex sentences appear around four-word utterances, around age 2 and up. Most of the types
used by age 4.
Individual differences in grammatical development
Children differ in both rate and course. Rate differences are the clearest.
Some use multiword utterances at 18 months; others begin 2 years. Some are unanalyzed wholes,
rote-learned; others use separate words from the beginning.
As discussed in other contexts, children have different approaches to acquiring syntax. This may
originate in differences in what children attend to, so how they perceive the speech they hear.
o Some attend more to syllables and phonemes, others more to the prosodic contour, prosodic
"tune." This is a holistic approach, top-down, resulting in perceiving unanalyzed chunks; these kids
may produce long utterances, but have little combinatorial ability. Others are more analytical,
bottom-up. The holistic kids eventually identify slots in phrases that can be occupied by other
lexical items, e.g. There's the xxx; Me got xxx; Wanna xxx. Most children use both bottom up and
top down strategies, but vary in reliance on one or other.
Measuring grammatical development
With increments in syntax mastery comes increasing utterance length. Length in morphemes is a
good index of the grammatical complexity of an utterance.
MLU better than age as an indicator of grammatical development. Once beyond 3.0 or 4.0, not so
good an indicator.
o I: MLU 1.01-1.99: begin to combine words
o II: MLU 2.00-2.49: begin to add grammatical morphemes to word combos.
o III: 2.50-2.99: begin to use different sentence modalities, e.g. negative and question forms.
o IV: 3.00 and up: begin to use complex sentences.
o V: marked by the appearance of new complex sentence forms. Age norms Table 6.1.
Development of comprehension of structured speech
Harder to study. If ask parents, they say their children understand "everything."
Strategies children use:
o Response strategies: enable them to respond to speech they only partially understand. Have
learned to respond to speech by doing something, so that when parent asks, e.g., "Why don't you
play with your blocks?", the child tending to blocks will indicate to parent child understood the
o If try to trip them up, may reveal lack of full comprehension. Child may be deriving sentence
meaning from knowledge about the world (single word comprehension) rather than knowledge of
syntactic structure. For example, if using word order strategy for "The swing bumps the kitty,"
doesn't work for the passive "The kitty was bumped by the swing." Or order-of-mention for complex
sentences, e.g. ‗John played before Mary sang" doesn't work for "John played after Mary sang."
Use probable event strategy, e.g. "The mouse was chased by the cat."
o Children do know more about their grammar than they reveal in spontaneous productions.
Comprehension of relational meanings in word combinations
Children at one word stage demonstrate an understanding of relational meanings beyond the
meanings of the words individually, e.g. "She's kissing the keys," and shown tapes simultaneously of a
woman kissing keys and holding up a ball, and a woman kissing a ball and holding up keys.
Comprehension of meaning in the structure of word combinations
At one word, children demonstrate understanding that meaning carried by word order.
o At 12-14 months, but not at 10 months, preferred to listen to sentences presented in normal word
order over scrambled word order, and specifically in their language; not necessarily understanding.
o At 16-18 months, distinguished word order "Where is Cookie Monster washing Big Bird" and
"Where is Big Bird washing Cookie Monster."
o 28 months, distinguish "Find Cookie Monster and Big Bird turning" and "Look, Cookie Monster is
turning Big Bird." Subject and object changes. Same vocabulary, but who is turning? At this stage,
only say things like "I watching cars."
o Same with grammatical development.
o So, the sequence of grammatical development occurring in comprehension is the same as
expression, but earlier. The duration of the lag is different for different children, thus an individual
child's comprehension skills are not a good predictor of that child's level of productive grammar.
o Also, grammar is accomplished quickly by very young children. Continue to learn about the
grammar, though, after age 5 years.
Children know the grammatical rules
As evidenced by
o Overregularization errors: they make mistakes in spontaneous speech that seem to be over-
applications of rules. Example in text: the child who said "amn't I" as in "I'm a good boy, amn't I" as
an over-application of the rule forming a tag question by inverting the subject and the form of the
verb to be. We say "aren't you" and "isn't he", but "amn't I" is an exception to the rule. In this case
the child has learned the rule but not the exception. Also often seen in overuse of grammatical
morphemes, e.g. "foots", "tooths", "goed." Evidence that they know the exception to the rule but
don't apply it, child making an irregular part of the language regular. Some overregularize more
o Also, children able to use novel forms, e.g. wugs, and blicked.