Lexical Acquisition and the Structure of the Mental Lexicon
Summary and Keywords
The words and word-parts children acquire at different stages offer insights into how the mental lexicon might be organized. Children first identify ‘words,’ recurring sequences of sounds, in the speech stream, attach some meaning to them, and, later, analyze such words further into parts, namely stems and affixes. These are the elements they store in memory in order to recognize them on subsequent occasions. They also serve as target models when children try to produce those words themselves. When they coin words, they make use of bare stems, combine certain stems with each other, and sometimes add affixes as well. The options they choose depend on how much they need to add to coin a new word, which familiar elements they can draw on, and how productive that option is in the language. Children’s uses of stems and affixes in coining new words also reveal that they must be relying on one representation in comprehension and a different representation in production. For comprehension, they need to store information about the acoustic properties of a word, taking into account different occasions, different speakers, and different dialects, not to mention second-language speakers. For production, they need to work out which articulatory plan to follow in order to reproduce the target word. And they take time to get their production of a word aligned with the representation they have stored for comprehension. In fact, there is a general asymmetry here, with comprehension being ahead of production for children, and also being far more extensive than production, for both children and adults. Finally, as children add more words to their repertoires, they organize and reorganize their vocabulary into semantic domains. In doing this, they make use of pragmatic directions from adults that help them link related words through a variety of semantic relations.
This article explores some implications of findings from first language acquisition for how we conceive of the mental lexicon. The focus will be on three topics: (a) words and parts of words stored in memory; (b) the general ‘gap’ between comprehension and production, what we can understand versus what we can say; and (c) how words within a semantic domain are linked to each other in meaning, and how those links get set up. All three topics have important implications for how we store items in the mental lexicon, and how we access and make use of them in processing language every day.
Children start on words in their acquisition of a first language very early. They identify recurring sequences in the speech stream and attend to the contexts where they hear these sequences (Jusczyk, 1997; Vihman, 1996). In taking their first steps towards language, children seize on recurring word forms, assign preliminary meanings to them, and store them in memory for future use. These uses are twofold: (a) they use stored word forms for accessing or looking up words they hear from others on subsequent occasions, in different contexts. And (b), they use them as templates or models for their own production, as well as models for when they repair their own versions until they align with the forms they have stored in memory. As children acquire more vocabulary, they also start to analyze parts of words: they distinguish stems and affixes, for example. In comprehension, they identify parts of word meanings in the constituent morphemes, and in production, they use familiar elements when they construct new words. I begin with a brief account of children’s general lexical acquisition and then turn to some of the insights the findings from acquisition provide into the nature of the mental lexicon.
1. Lexical Acquisition
Children give their first evidence of understanding one or two words as early as six to nine months old, but more generally display some preliminary understanding of their first few words at around 9 to 12 months (see Bergelson & Swingley, 2015; Benedict, 1979; Oviatt, 1980). They also recognize the communicative function of words at around the same time (Vouloumanos, Martin, & Onishi, 2014). Their earliest attempts at word production follow soon after, around 12 to 14 months, but these first attempts often fail because their words are unrecognizable. Indeed, it may take young children several weeks of trying before they can make their version of a word be understood (e.g., Scollon, 1976; Dromi, 1987). At 11 months, the median number of words children understand, as measured by the Communicative Development Inventory, is 54; this increases to 169 words at 16 months (with around 340 words at the 90th percentile). The number of words produced is much lower, with fewer than 10 words at 12 months, increasing to a median value of 40 words at 16 months. By 30 months (2;6), the median score rises to 573, a tenfold increase in production vocabulary from 1;4 to 2;6 (Fenson, Dale, Reznick, Bates, Thal, & Pethick, 1994).
Children also make consistent use of gestures such as pointing, holding out (showing), and reaching from 8–10 months on. Adults respond differentially to these gestures: They offer labels in response to points, they offer help in doing something in response to showing gestures, and they try to discern what children want when they make reaching gestures (e.g., Olson & Masur, 2011, 2013; Kelly, 2011, 2014). Children combine these gestures with single words, and so serve to indicate where something is at the same time that the child labels the object of interest, as in ‘POINT + cat’, with a general meaning of ‘there’s a/the cat’. These early gesture-and-word combinations are precursors to word–word combinations in children’s early utterances (e.g., Capirci, Iverson, Pizzuto, & Volterra, 1996).
When children hear a new word, they make inferences about its possible meaning based on joint attention, physical copresence, and conversational copresence. That is, they infer that the new word is relevant to whatever the child and adult are currently attending to, and therefore take into account both the physical setting they are in and any familiar words in what the adult says on that occasion. These factors guide children’s pragmatic inferences about possible and probable meanings in context. While children are exposed to many unfamiliar words on the fly and must make inferences right away, sometimes based initially on minimal information about the probable meaning, they also hear some explicit offers of new words. Adults tend to flag these as new by using introductory phrases like ‘This is a —‘ or ‘That’s called a —,’ and they often add additional information about the referent, identifying its class membership, or salient parts and properties, for example (Clark, 1998; Clark & Wong, 2002; Clark, 2007; Clark & Estigarribia, 2011). And children give evidence of attending to such offers in their uptake: they repeat and acknowledge new words, and often take up and elaborate on any added information adults provide (Clark, 2002; 2010).
As children add words to their repertoire, they need to store them in memory so they will recognize them when they hear them on future occasions. Once stored, each word form can also serve as a template to aim for as they try to produce that word themselves. What children store in memory is based on the adult production of each word. As they hear more instances of adult uses, they consolidate the form and minor variants of that form heard in speech from different people. At the same time, they gradually elaborate and consolidate its meaning as they compare instances of adult use across different contexts (Clark, 2017a). These words in memory then become retrievable for children, both for recognition when they hear others speaking, and for production when they are planning what to say themselves. In effect, for a word to be acquired requires that (a) it be encoded in memory, and (b) its form and meaning be consolidated. It can then be retrieved as needed and be used as a template for production (Clark, 1993; 2016).
2. Analyzing Word Forms
As children acquire language, they demonstrate considerable awareness of the elements used as building blocks in words. They make spontaneous analyses of word forms, commenting on the constituents they recognize ‘inside’ more complex words such as compounds as in (1)–(5):
In addition to comments like these on constituent morphemes in compound words, they also show they are aware of smaller elements, detecting rhymes, and sounding out short words (even saying them backwards, as in a word like CAT /kæt/ transformed to /t– æ– k–/ then /tæk/) (e.g., Fox & Routh, 1975; Clark, 1978, 2015; Slobin, 1978).
Children also readily offer interpretations of novel words built on familiar stems. For example, when presented with a picture of a face and an introduction like: ‘This man is an opener. What does an opener do?’, children, from age 3 on, provide such interpretations as He opens, He opens things, or Opens things (Clark & Hecht, 1982). Between 3;0 and 3;8, children acquiring English can extract the verb stem from such novel words and give an appropriate gloss for unfamiliar agentive nouns 93% of the time. (Children acquiring Icelandic do equally well, giving appropriate glosses 86% of the time.) And they do nearly as well in analyzing novel instrument nouns. In response to such introductions as: ‘This is a breaker. What’s a breaker for?’ children offer glosses like It breaks, It breaks things, for example, 79% of the time for English forms, 76% of the time for Icelandic ones (Clark & Hecht, 1982; Mulford, 1983). Children acquiring Hebrew offer glosses when asked to just as readily, and at age 3 provide appropriate glosses of both agentive and instrumental noun patterns about half the time. But, as in English, by age 4 to 5, most children are at ceiling on such analyses (Clark & Berman, 1984).
These studies strongly suggest that children as young as 3, who are presented with unfamiliar words built on a familiar stem and an affix, for example, are capable of constructing a gloss or interpretation by doing the following:
(a) strip off any agentive or instrumental suffix
(b) extract the verb stem and its meaning
(c) add a lexical or pronominal subject
(d) add any inflectional ending needed on the verb
(e) and add a lexical object of some kind
In short, children recognize and extract familiar elements (here, the verb stems) in unfamiliar words. And, after extracting a familiar element, they can use it in constructing their interpretation by taking the verb, adding a subject, inflecting the verb for third person agreement (they typically give generic definitions), and on occasion adding a direct object, often the general word things. In effect, they apply the same skills in analyzing unfamiliar words that they apply in processing adult speech in general. They draw on units stored in memory, either whole words consisting of a single morpheme and its meaning, or complex words consisting of several different morphemes, where they may or may not be able to analyze the latter into constituent parts, namely stems and stem combinations, suffixes and prefixes. By drawing on the elements they have stored, children are able to make inferences about possible meanings of new words in context. Adults do this too, looking first for any familiar elements in such words as tachistoscope, widdershins (and diesel), or eleemosynary, and only when that strategy fails, do we ask for the meaning, or else head to a dictionary.
3. Producing Novel Words
The ability to recognize familiar word stems as well as an increasing number of suffixes is a necessary prerequisite for coining words for oneself. Children produce a variety of coinages from early on, including denominal verbs, denominal and deverbal adjectives, compound nouns, and derived nouns. Some typical examples of denominal verbs coined by children are given in Table 1.
Table 1: New Denominal Verbs—Typical instances
Denominal verb formation is particularly productive in English, and children produce such verbs from as early as age 2 (e.g., Clark, 1982; 1993). In forming such verbs, all children need do is take the relevant noun stem and add verb inflections, so the cost here of constructing a new word is minimal. Children also coin denominal and deverbal adjectives, as shown in Table 2. These are also low-cost options, although not as productive as the formation of denominal verbs in English.
Table 2: New Denominal and Deverbal Adjectives—Typical instances
Analysis of all the novel adjectival forms produced by the child D showed that he consistently opted for a noun-stem+y form when talking about inherent or permanent properties, but a verb-stem+ed when talking about a temporary property or state of affairs (Clark, 2001). Although English adjectives as a whole are not consistent with this division of labor, D started out with this analysis, which allowed for two distinct meanings to be assigned to the two suffixes, with ‑y for inherent properties, and ‑ed for temporary ones, and he maintained it for some time in his adjectival coinages.
Another type of coinage that emerges early in English, and in other Germanic languages, is compounding with root compounds, compound nouns made up, generally, of a modifier noun plus head noun, as in the typical examples in Table 3.
Table 3: New Compound Nouns—Typical instances
Young children coin compound nouns to capture contrasts among subcategories, as in chimney-smoke vs. car-smoke, or Dalmatian-dog vs. boxer-dog. They sometimes appear in pairs that mark the contrast overtly, sometimes as singletons with the contrast inferable from context (see Clark, Gelman, & Lane, 1985; Gelman, Wilcox, & Clark, 1989; Berman & Clark, 1989). Synthetic compounds, as shown in (6) and (7), are rather rarer among children’s spontaneous coinages, only starting to emerge around age 3 to 4.
To produce synthetic compounds, children combine a verb and a noun root, in the order noun + verb, and add a suffix to the whole. But managing this construction is fraught with errors in production: children retain the predicate order of verb + noun and use only that combination (a build-wall); they add the suffix appropriate to the verb but keep the verb as the first element in the compound (a builder-wall), and, occasionally, they add the suffix to the second element, going from build-wall to build-waller. It is only around age 5 to 6 that children reliably construct synthetic compounds like wall-builder for ‘someone who builds walls’ in English (Clark, Hecht, & Mulford, 1986). The ‑er suffix here, of course, is the agentive ‑er that children are able to gloss successfully from around age 3.
When do children start to produce derivational suffixes? These start to emerge in their spontaneous coinages in the middle of their third year, with novel agentive forms like those in (8)–(13):
Novel instrumental forms like those in (14)–(18) are not as common in spontaneous speech:
In essence, children acquiring English identify one agentive suffix and, starting between age 3 and 4, use that in their spontaneous speech to coin new agent nouns. (Before that, they tend to rely on root compounds.) The suffix they choose is then extended to novel instrument nouns as well. The same pattern appears in the coinages constructed by children acquiring Icelandic, German, and Swedish among Germanic languages, and also by children acquiring French, Italian, and Hebrew (see Clark, 1993).
Children display the same tendencies in elicitation tasks where they are asked to come up with a new word when they are given an agentive definition like ‘This is a picture of a boy who kicks things. What can we call him? He’s a —.’ Or an instrumental definition like ‘This is something for throwing buttons. What could we call it? It’s a —.’ Children are consistent in choosing the most productive of the available agentive suffixes in their language, and then, typically, extending it to novel instruments as well. This can be seen in Tables 4 and 5, which summarize some of the data from experimental elicitation studies of English and Hebrew (Clark & Hecht, 1982; Clark & Berman, 1984). Table 4 displays the children’s preference for ‑er as the dominant suffix in English for forming new agent and instrument nouns. When the children didn’t use ‑er, they generally relied on root compounds (e.g., open-man, push-thing) and at times opted for established words and even, for agents, the occasional proper name. They did not use the less productive agentive suffixes, ‑ist and ‑ian at this stage. These emerge only later in children’s usage (see Clark & Cohen, 1984).
Table 4: Percentage of -er suffixes in novel agent and instrument nouns in English
3;0 – 3;8
3;9 – 4;5
4;6 – 5;2
5;3 – 6;0
In Hebrew, children also choose the more productive of the main options for coining agent nouns, and typically rely on the CaCCan pattern, with its agentive suffix ‑an, as in rakdan ‘dancer’ or saxkan ‘actor’. And, as in other languages, children acquiring Hebrew extend this pattern to instrument nouns, although normatively, this noun pattern is only for agents, not instruments. However, adults also overextend this pattern in colloquial Hebrew (see Berman, 1987). Otherwise, instrument nouns appear in one of two main instrument patterns, maCCeC or maCCeCa. The degree to which children of different ages rely on the suffix ‑an (as in CaCCan forms) is shown in Table 5. And children acquiring Hebrew use this pattern almost as often for novel instrument nouns as they do for novel agents.
Table 5: Percentage of CaCCan forms in novel agent and instrument nouns in Hebrew
3;0 – 3;11
4;2 – 4;9
5;0 – 5;9
7;3 – 8;0
11;0 – 12;0
The findings for both English and Hebrew here show that children are sensitive to the word-shapes that are the most productive in current usage for the expression of agency in both languages. Indeed, in studies of word-formation more generally, such productivity appears to determine what word-shapes children will pick up first when they begin to coin words with particular kinds of meaning. Another factor that appears to guide early acquisition is simplicity of form such that earlier word shapes used in coinages require few or no changes in the shape of the source word (as when children acquiring English go from a noun to a verb, e.g., I broomed her). Hence, for example, children’s reliance on denominal verbs in languages where this is a productive option. This reliance on simplicity would also account for the early uses of root compounds, again in languages where this is a productive option. In addition to productivity and simplicity, children also rely on transparency: they consistently make use of stems and affixes for which they have already mastered the meanings (Clark, 1993). For example, in English, for agency, they first rely on ‑er, and only later, from age 4 to 5 on do they start to produce ‑ist (linked first to players of different instruments, as in flutist and the innovative trumpetist), and only later still ‑ian (Clark & Cohen, 1984).
Early on, around age 2 to 3, children also make some use of the English diminutive ‑ie/‑y (as in such child forms as catty, chair-y, ball-y), but reliance on diminutive forms is much more extensive in Slavic than in Germanic languages (e.g., Kempe, Brooks, Mironova, & Fedorova, 2003; Seva, Kempe, Brooks, Mironova, Pershukova, & Fedorova, 2007). For instruments, children acquiring English rely on ‑er and ‑Ø (as in the established instrument terms ruler and drill). By age 5 to 6, they have also begun to use further suffixes like ‑ness and ‑ly, as well as the reversal prefix un‑ (Clark, Carpenter, & Deutsch, 1995). However, the order in which affixes for expressing particular types of meaning are acquired and emerge in children’s coinages varies by language. In Romance languages, for example, children acquire a large range of suffixes earlier, in part because in those languages derivational forms are considerably more productive than compounding (Clark, 2014). Finally, there is little or no research in acquisition on other processes in word formation such as prefixation, reduplication, or infixation.
4. Comprehension versus Production
Comprehension is ahead of production from the start. And it remains ahead, in that it is much more extensive than production. In comprehension, both adults and children need to be able to deal with different speakers, different social registers, different dialects (with differences in both pronunciation and vocabulary), infrequent words, and varieties of the language from different eras (e.g., 15th- or 17th-century English compared to some variety of 21st-century English). Yet as speakers, in production, we typically master just one dialect, the one in use in the speech community we grow up in. So speakers’ production of language is typically matched only to a subset of all they can understand. This asymmetry holds for second language learning too: speakers typically understand more than they can say in a second language.
Consider the differences involved in the two processes. In comprehension, we hear acoustic sequences, segment them into words, and make use of these in accessing and retrieving from memory the relevant meaning associated with each lexical item. But in production, we must plan what to say, retrieve the relevant lexical items from memory, put them into an appropriate construction, add any affixes required, and then produce the resultant utterance using the appropriate articulatory plan (see Levelt, 1989).
This gap between comprehension and production shows up at numerous points during acquisition. Children consistently do better in comprehension tasks than in production (see further Clark & Hecht, 1983). This is also true for the acquisition of word-formational patterns in the lexicon. For example, when children are given novel agent and instrument nouns in English and asked to gloss their meanings, they do very well in this as young as age 3, and distinguish stems from affixes—an opener is ‘someone that opens / that opens things,’ a builder is ‘a man that builds,’ and so on, readily identifying the verb stem inside the agent noun. But when they are asked to construct agent and instrument nouns themselves, they rely initially on root compounds of the form move-lady or push-man (in response to such questions as ‘What could you call a woman who moves things?’ or ‘What could you call someone who pushes things?’). But these root compounds for agents, often formed at this stage by combining a verb and a noun stem, are not actually productive in English, unlike children’s early noun–noun root compounds (e.g., Clark, Gelman, & Lane, 1985). Compounding, though, tends to be acquired in production before any derivational options in English, hence before children master suffixes like the agentive or instrumental ‑er (Clark & Hecht, 1982). Children display similar asymmetries in acquiring the relevant patterns in Hebrew, displaying the ability to interpret new agent and instrument nouns well before they manage to produce them themselves (Clark & Berman, 1984).
This gap between comprehension and production sometimes results in both quantitative and qualitative differences. Consider synthetic compounds in English, of the general type clock-mender for ‘someone who mends clocks.’ In one study, we checked on children’s comprehension by asking them to explain what a particular person did, in response, say, to ‘I have a picture of someone here: he’s a button-thrower. What do you think a button-thrower is?’ Four- and five-year-olds respond with interpretations like ‘throws buttons,’ ‘he throws buttons,’ and ‘he throws things,’ that show they are able to unpack the synthetic compound into the relevant verb, noun, and affix constituents. In short, they demonstrate good comprehension of this word-shape. Furthermore, at age 4, they identify the V+er in such compounds as the head 66% of the time, and by age 5 do so 80% of the time, as revealed by their interpretations.
But in production, things look rather different: In response to ‘What could you call someone who pulls wagons?’, the same children are likely to respond with such forms as a WAGON-man, a PULL-man, a PULLER-man, and, only from a few of the older children (aged 5 and more) a WAGON-puller (see Clark, 1984; Clark, Hecht, & Mulford, 1986; Clark & Barron, 1988). But acceptable forms in production would be: a WAGON-man (N+N), a puller (V+er), a wagoner (N+er), and, the productive adult preference, a WAGON-puller.
The commonest forms children produce before they arrive at the ‘target’ adult agent form, namely N+V+er, include the ungrammatical word shapes in (19):
All these forms were produced with the appropriate compound stress: primary stress on the (initial) modifier slot and tertiary stress on the head slot in the compound. In some cases, the children tried to make clear they were talking about an agent by adding a classifier-like noun, usually ‑man as the head (in 20b, d, f). Notice though that all these patterns retained predicate-like ordering of the verb and direct object. The word order mismatch between predicate ordering and compound ordering of the verb stem and direct object appears to be what causes English-speaking children the most difficulty here. And at ages 4 and 5, they produce this order on average in about 40% their compounds. At the same time, they are aware that such affixes as ‑ing and ‑er attach to the verb stem. Notice that in languages where the word order in predicates and compounds is the same, children simply never make errors of this kind (see Clark & Berman, 1987).
If four- and five-year-old children hear erroneous compound forms of the type they often produce, how do they interpret them? To answer this question, Clark (1984) took each child’s dominant error-type in production and constructed a set of compounds based on the same pattern for each child, then asked the children to provide glosses for these forms in a follow-up test of their comprehension. (These children had earlier shown that they consistently took the first element in synthetic compounds to be a modifier and the second, rightmost element to be the head.) In the follow-up of their interpretations of the main erroneous forms they produced, children showed that they consistently interpreted the rightmost element in each compound type as the head. The result, for the three tasks: the initial comprehension task, the production task, and the second comprehension task, is summarized in (20).
In short, children start out with a representation for comprehension in which they have identified the modifier and head in synthetic compounds. They use this information in their interpretations of what a novel compound could mean (20i). In production, though, they start out by adhering to the ordering of verb and noun that is familiar from predicate ordering in English (V before N), and rely on this ordering as they construct compounds in production (20ii). The next step, which can take several years, requires that they align these mismatched representations. In effect, they have to adjust their schemas for production until they match their representations for comprehension. This evidently takes some time (Clark, 1984; Clark, Hecht, & Mulford, 1986).
To summarize, the general asymmetry of comprehension and production is sometimes particularly evident in children’s novel word formation. They can interpret and gloss novel word forms they encounter, but can’t always produce the relevant form. The ordering errors they make appear to be attributable to the mismatch in Verb and Noun (Direct Object) order in predicates compared to synthetic compounds with agent meanings. The findings here offer striking evidence for distinct representations at work in comprehension versus production. This suggests that we do not store just one representation for a unit of meaning, but at least two—one for comprehension and one for production (see further Clark, 2003).
5. Semantic Domains and Semantic Relations
The lexicon doesn’t just consist of unconnected words. Rather it contains all kinds of semantic domains that group particular sets of words together by means of various semantic relations. Children’s earliest domains, gradually enlarged as they add more words that are relevant, depend on such semantic relations as is-a-kind-of (‘a seal is a kind of animal’), has-a or is-a-part-of (‘a rabbit has a white tail,’ ‘the steering wheel is part of the tractor’), is-for (‘that knife is for cutting chicken’), and is-made-of (‘balls are made of rubber’), as well as information about characteristic sounds, motion, habitat, ontogenesis, and other types of information. Adults often supply such relations as they link new words to familiar ones in talk about the relevant domain. Effectively, they offer pragmatic directions in this way about how to use the new terms, by providing added information about the referent object, event, or relation (e.g., Clark, 1998; Clark & Wong, 2002; Clark & Estigarribia, 2011).
Some semantic domains or fields are loosely structured sets of nouns and verbs used for a specific activity such as eating and drinking, or gardening, or sailing. Others consist in part of hierarchically structured terms, with anywhere from two to seven levels, as in various natural kind terms for plants and animals (e.g., plant, tree, pine, White Pine, Northern White pine; or animal, mammal, horse, Arabian). Such domains also contain terms for the activities linked to the objects and their roles (see Lyons, 1977; Lehrer & Kittay, 1992). The more words children acquire in a domain, the better that domain becomes organized. As result, children become better able to retrieve the words they already know, and to add new words and connect them to any information relevant both to the domain as a whole and to the particular categories they have just acquired words for.
Consider the domain of space and spatial terms, a domain children must invoke whenever they talk about relations and motion in space. In English, children must learn how to talk about such topological relations as containment, support, and proximity for talking about (static) relations in space. But they must also learn terms for actions of placement (objects that are in, on, under, or next to some reference point or landmark), of joining (linking pop beads, say) and separating (removing lids), holding and carrying, as well as keeping track of manner, motion, and path in talking about movement in space (Bowerman & Choi, 2003; Clark, 2017a). In learning to talk about such spatial relations, children acquiring English need to master the preposition in for talk about containment of all kinds, and on for support from an underlying surface. They also need to take into account the verb used, but this tends to come later with general purpose put, then put in, put on, take out, take off, and so on. The spatial relation, though, is generally clear from the preposition alone. In Korean, children must start with verbs since these are what capture the spatial relation. For containment, they also have to distinguish close-fit containment, where the object being placed fits tightly into the container (a cassette in a cassette case, the top on a pen, a triangle in the triangular hole in a shape box), from loose-fit containment (an apple in a bowl, toys in a box) (Choi & Bowerman, 1991). As children talk about joining and separating objects, they must also attend to properties of the objects and their juxtapositions in space. Korean children must learn specific verbs for removing objects from containers and supporting surfaces, for example, where children acquiring English often rely on open, or even undo, before mastering the conventional take combined with a negative particle, as in take out, take off, take away (Griffiths & Atkinson, 1978; Clark, Carpenter, & Deutsch, 1995; Bowerman & Choi, 2003).
As children begin to talk about motion in space, they also have to attend to whether the verbs they are acquiring are satellite-framed, as in English, where the path of motion is indicated by the particles that follow the verb (e.g., go in, go out, go along), or verb-framed, as in Spanish, where information about the path in included in the verb itself (e.g., bajar ‘go down’, subir ‘go up’, entrar ‘go in’). Satellite-framed languages typically include ‘manner of motion’ in the verb, as in English stroll, wander, jog, stumble, etc., but this information has to be added separately in verb-framed languages since it is not generally included in the meaning of motion verbs. In addition to the verbs to this domain, children also have to work on the syntax and semantics of sources and goals when they specify motion.
How do children set up such semantic domains? They need to accumulate information about the referents of words for objects, actions, events, and relations in the speech they hear from more expert speakers. Besides unfamiliar words that children must make inferences about on the fly, they also hear explicit offers of new words. And with these offers, adults also provide pragmatic directions for their use—in the form of information about the referents of the words (Clark, 2007, 2010; Clark & Grossman, 1998). In doing this, adults provide information about set membership or inclusion, parts and properties, characteristic motion and sound, functions, ontogeny, and habitat, as well as making appeals to memory for events the child has experienced that are relevant to the information being provided (Clark, 1998; Clark & Wong, 2002; Clark, 2010; Clark & Estigarribia, 2011). In short, adults embed new words in the appropriate domain, and identify other, familiar, words that the new ones are being linked to.
Pragmatic directions that link new words to words already known to the child offer one source of information about how word meanings are connected. This helps children organize what they know about particular domains, and organize the words they already know for each domain. At the same time, there are many everyday words that crosscut any domain-based organization of the lexicon since they apply across multiple domains. This suggests that there must be different levels of lexical organization, depending on the degree of expertise the speaker has about any particular domain. This in turn accounts for why one speaker may have a larger vocabulary for sailing, another for mountain climbing, another for medicine, another for neuroscience, and another for gardening. To become an expert, whether on cars or dinosaurs, or on any other domain, children and adults need the relevant words—the vocabulary for that domain. So, as children add to their early vocabularies, they add both general terms, terms that they will likely rely on every day, and also, as they get older, terms specific to particular domains that reflect their accumulation of knowledge, even expertise, in those domains. Their initial organization of each domain depends on where they begin—the first words they acquire, and on the amount of input they receive and accumulate (see further Clark, 2017a).
6. Some Implications
These findings have a number of implications for how both children and adults process the language they hear and produce. Consider word access and retrieval: Children and adults draw on terms already stored in memory both for comprehension and production. In comprehension, they look up words they hear in order to arrive at an interpretation. In production, they search for words that capture the intention they wish to express as they are planning their utterances. For both processes, speakers need to be able to access and retrieve words, parts of words, and also chunks or phrases stored in their mental lexicon.
To build up a repertoire of words and phrases, children need extensive exposure to language, especially in the first three years, when this both strengthens young children’s processing skills, particularly in recognizing familiar words, and builds vocabulary. The more adults talk to their children, the faster their children get in processing familiar words. For example, after hearing phrases like ‘Where is the —?’ while being shown two referent pictures on a screen, one- and two-year-olds look faster with age to the picture of the referent mentioned. And as they get faster at this task, they also add more words to their vocabulary. That is, there are high correlations between amount of speech heard, processing speed, and new vocabulary acquisition in the first three years (e.g., Weisleder & Fernald, 2013).
Findings like these have built on earlier work showing that the amount of speech addressed to young children up to age 3 or 4 is highly correlated with later language development and with success in elementary school (e.g., Hart & Risley, 1995, Walker, Greenwood, Hart, & Carta, 1994). At the same time, there is considerable variation between social classes, with upper- and middle-class parents talking to their children considerably more than lower-class parents do. And within social classes, there is also variation from one family to another, both in the amount of talk to children and in the sophistication of the words used. Both make a difference to later outcomes (see Weisleder & Fernald, 2013; Weizman & Snow, 2001; Walker, Greenwood, Hart, & Carta, 1994). A further factor in expanding children’s experience with language is the range of speakers they are exposed to and are expected to respond to. Children with greater exposure early to speakers outside their own family, in daycare for example, tend to do better on a variety of language tests: they are used to telling parents about what they did all day and to telling other adults what they want to do and what they like (or don’t like). They get more practice in using language with others (see Marcos, Salazar Orvig, Bernicot, Guidetti, Hudelot, & Préneron, 2004; Hoff, 2006; Clark, 2017b). Moreover, with more interactions using language, children not only get practice but also receive extensive feedback (e.g., Chouinard & Clark, 2003; Farrar, 1992; Saxton, Backley, & Gallaway, 2005; Strapp, Bleakney, Helmick, & Tonkovich, 2008). This includes feedback on word choice, where adults typically offer both embedded corrections, in which they substitute the conventional term for the child’s erroneous choice, and side sequences where they query the child’s choice by offering offer an alternative term. Such feedback supplements the kinds of pragmatic directions adults provide when they talk about the properties of the referents of new words. At the same time, general exposure to language use in conversation, in joint activities, and in reading, all add to what children know about words, word meanings, and the parts of words that contribute to those meanings.
The asymmetries observed in children’s acquisition of word formation patterns mirror a general feature of language at large. Speakers understand much more than they say, in both first and second languages. This asymmetry allows them to understand different speakers, from different dialects as well as many accented second-language speakers, and also allows them access to older varieties of their language(s). Early exposure to different speakers and dialects makes young children more proficient in understanding more of the variants they are exposed to (see Clark, 2016, 2017b). Notice, too, that there is no communicative need for children, or adults, to be able to produce a range of such variations: they can make themselves understood without it.
For comprehension, children, like adults, store word forms or templates in memory for use in recognizing words they hear and in identifying parts of words. This is critical when they attempt to analyze new words by identifying any parts they recognize—noun and verb stems, for example—and extracting affixes. Once they have represented these for comprehension, based on their appearance in familiar words, they can draw on them in interpreting new words (e.g., Clark 1984; Clark & Berman, 1987). They can also draw on them in production: affixes that are transparent in meaning become available, along with familiar nouns and verbs, as elements that they can make use of when they coin new words with new meanings (Clark, 1993).
Another factor that plays a role in language processing is frequency. But frequency itself is a product of age-of-acquisition and of current productivity in a language. First, words acquired earlier tend to be more frequent in the language and are easier to find in memory. This affects how fast children (and adults) can retrieve the meanings of words from memory as they read and do other kinds of language-based tasks (e.g., Zevin & Seidenberg, 2002). Second, age of acquisition and frequency also predict has fast people can label pictures as well as retrieve word meanings (e.g., Carroll & White, 1973; Morrison, Ellis, & Quinlan, 1992; Ellis & Morrison, 1998).
But are the most frequent nouns and verbs, along with affixes, which are the most available items, also the most productive options, for constructing new words? Here, the findings from acquisition show that frequency per se has to be distinguished from current productivity in a language. Productivity offers a measure of what speakers favor when choosing how to construct new words. This contrasts with frequency, which is generally a measure of past productivity as recorded in dictionaries (Clark & Berman, 1984; Berman, 1987). Children are sensitive to the relative productivity of particular word-formational options: they pick up on the most productive option for expressing a particular meaning in their early coinages, and only later acquire rather less productive options. In fact, what is productive in word formation in a language tends to change over time, so that while speakers of French in one era favored one suffix for new agent nouns (the suffix ‑ien) in the 19th century, a century later speakers favored a different suffix, ‑eur, instead (e.g., Dubois, 1962; Guilbert, 1975). Similar shifts have been documented in Polish (e.g., Chmura-Klekotowa, 1970;). Changes in productivity take place in the usage of active speakers of a language, and so cannot be assessed from written sources. Productivity contributes directly to frequency in word formation.
The units used in word formation must be units stored in memory as phonological sequences associated with particular meanings. But to what extent do children, and adults, store roots and affixes, for example, both in conventional combinations and as separate units available for use in new combinations when coining words? Do the representations for phonological sequences coincide with those for morphological units and for ready-made, conventional words? Data from acquisition here offer only partial answers so far: Children do make use of subparts of words, they distinguish stems from affixes, and they attend to how productive different affixes are in current adult usage. Like adults, they also come to understand and use extensions of conventional terms in metaphors, for example (e.g., Duvignau, 2007; Winner, 1997). But there are still many issues that remain to be explored with respect to the nature of the various representations people appear to rely on as they use language (Gentner & Goldin-Meadow, 2003).
Lexical acquisition offers insight into several aspects of the mental lexicon. First of all, it reveals which units of language young speakers focus on, how early they can identify the noun and verb stems they make use of in early word coinages, as well as the first affixes they identify and use. Secondly, findings from lexical acquisition offer detailed evidence that children set up distinct representations for comprehension and production and that they clearly make use of both representations for a word as they progress from their earliest productions (often unrecognizable) to more adultlike forms. This offers further motivation for assuming a more complex system of representations that depends on how each one is used. Third, data from acquisition offer insight in how children organize and at times reorganize the lexicon as they acquire more words and set up semantic domains.
Adams, V. (1973). An introduction to modern English word-formation. London: Longmans.Find this resource:
Bloom, P. (2000). How children learn the meanings of words. Cambridge, MA: MIT Press.Find this resource:
Clark, E. V. (1993). The lexicon in acquisition. Cambridge, U.K.: Cambridge University Press.Find this resource:
Clark, E. V. (2016). First language acquisition (3d ed.). Cambridge, U.K.: Cambridge University Press.Find this resource:
Cruse, A. (2011). Meaning in language (3d ed.). Oxford: Oxford University Press.Find this resource:
Lehrer, A. (1974). Semantic fields and lexical structure. Amsterdam: North-Holland.Find this resource:
Lyons, J. (1977). Semantics (2 vols.). Cambridge, U.K.: Cambridge University Press.Find this resource:
Marchand, H. (1969). English word-formation (2d ed.). München: C. H. Beck.Find this resource:
Benedict, H. (1979). Early lexical development: Comprehension and production. Journal of Child Language, 6, 183–200.Find this resource:
Bergelson, E., & Swingley, D. (2015). Early word comprehension in infants: Replication and extension. Language Learning and Development, 11, 360–380.Find this resource:
Berman, R. A. (1987). Productivity in the lexicon: New-word formation in Modern Hebrew. Folia Linguistica, 21, 425–461.Find this resource:
Berman, R. A., & Clark, E. V. (1989). Learning to use compounds for contrast: Data from Hebrew. First Language, 9, 247–270.Find this resource:
Bowerman, M., & Choi, S. (2003). Space under construction: Language-specific spatial categorization in first language acquisition. In D. Gentner & S. Goldin-Meadow (Eds.), Language in mind (pp. 387–427). Cambridge, MA: MIT.Find this resource:
Capirci, O., Iverson, J. M., Pizzuto, E., & Volterra, V. (1996). Gestures and words during the transition to two-word speech. Journal of Child Language, 23, 645–673.Find this resource:
Carroll, J. B., & White, M. N. (1973). Word frequency and age of acquisition as determiners of picture-naming latency. The Quarterly Journal of Experimental Psychology, 25, 85–95.Find this resource:
Chmura-Klekotowa, M. (1970). Odbicie tendencji słowotwórczych języka polskiego w neologizmach dzieci. Prace Filologiczne, 20, 153–159.Find this resource:
Choi, S., & Bowerman, M. (1991). Learning to express motion events in English and Korean: The influence of language-specific lexicalization patterns. Cognition, 41, 83–121.Find this resource:
Chouinard, M. M., & Clark, E. V. (2003). Adult reformulations of child errors as negative evidence. Journal of Child Language, 30, 637–669.Find this resource:
Clark, E. V. (1978). Awareness of language: Some evidence from what children say and do. In A. Sinclair, R. Jarvella, & W. J. M. Levelt (Eds.), The child’s conception of language (pp. 17–43). Berlin: Springer-verlag.Find this resource:
Clark, E. V. (1982). The young word-maker: A case study of innovation in the child’s lexicon. In E. Wanner & L. R. Gleitman (Eds.), Language acquisition: The state of the art (pp. 390–425). Cambridge, U.K.: Cambridge University Press.Find this resource:
Clark, E. V. (1984). Acquiring compounds. In ESCOL ’84: Proceedings of the First Eastern States Conference on Linguistics (pp. 181–190). Columbus, Ohio: Ohio State University, Department of Linguistics.Find this resource:
Clark, E. V. (1993). The lexicon in acquisition. Cambridge, U.K.: Cambridge University Press.Find this resource:
Clark, E. V. (1998). Lexical structure and pragmatic directions in acquisition. In M. C. Gruber, D. Higgins, K. S. Olson, & T. Wysocki (Eds.), Papers from the 34th meeting of the Chicago Linguistic Society 1998: The panels (pp. 437–446). Chicago: Chicago Linguistic Society.Find this resource:
Clark, E. V. (2001). Emergent categories in first language acquisition. In M. Bowerman & S. C. Levinson (Eds.), Language acquisition and conceptual development (pp. 379–405). Cambridge, U.K.: Cambridge University Press.Find this resource:
Clark, E. V. (2002). Making use of pragmatic inferences in the acquisition of meaning. In D. Beaver, S. Kaufmann, B. Z. Clark, & L. Casillas (Eds.), The construction of meaning (pp. 45–58). Stanford, CA: CSLI Publications.Find this resource:
Clark, E. V. (2003). Languages and representations. In D. Gentner & S. E. Goldin-Meadow (Eds.), Language in mind (pp. 17–24). Cambridge, MA: MIT/BradfordFind this resource:
Clark, E. V. (2007). Young children’s uptake of new words in conversation. Language in Society, 36, 157–182.Find this resource:
Clark, E. V. (2010). Adult offer, word-class, and child uptake in early lexical acquisition. First Language, 30, 250–269.Find this resource:
Clark, E. V. (2014). The acquisition of derivation. In R. Lieber & P. Stekauer (Eds.), Handbook of derivation (pp. 424–439). Oxford: Oxford University Press.Find this resource:
Clark, E. V. (2015). Linguistic units in language acquisition. In H. Narrog & B. Heine (Eds.), Oxford handbook of linguistic analysis (2d ed.) (pp. 27–42). Oxford: Oxford University Press.Find this resource:
Clark, E. V. (2016). First language acquisition (3d ed.). Cambridge, U.K.: Cambridge University Press.Find this resource:
Clark, E. V. (2017a). Acquiring meanings and meaning domains. In K. Syrett & S. Arunachalam (Eds.), Semantics in language acquisition. Amsterdam: John Benjamins.Find this resource:
Clark, E. V. (2017b). Conversational partners and common ground: Variation contributes to language acquisition. In M. Hickmann, E. Veneziano, & H. Jisa (Eds.), Sources of variation in first language acquisition: Languages, contexts, learners. Amsterdam: John Benjamins.Find this resource:
Clark, E. V., & Barron, B. J. S. (1988). A thrower-button or a button-thrower? Children’s judgments of grammatical and ungrammatical compound nouns. Linguistics, 26, 3–20.Find this resource:
Clark, E. V., & Berman, R. A. (1984). Structure and use in the acquisition of word formation. Language, 60, 542–590.Find this resource:
Clark, E. V., & Berman, R. A. (1987). Types of linguistic knowledge: Interpreting and producing compound nouns. Journal of Child Language, 14, 547–567.Find this resource:
Clark, E. V., Carpenter, K. L., & Deutsch, W. (1995). Reference states and reversals: Undoing actions with verbs. Journal of Child Language, 22, 633–662.Find this resource:
Clark, E. V., & Cohen, S. R. (1984). Productivity and memory for newly formed words. Journal of Child Language, 11, 611–625.Find this resource:
Clark, E. V., & Estigarribia, B. (2011). Using speech and gesture to inform young children about unfamiliar word meanings. Gesture, 11, 1–23.Find this resource:
Clark, E. V., Gelman, S. A., & Lane, N. M. (1985). Compound nouns and category structure in young children. Child Development, 56, 84–94.Find this resource:
Clark, E. V., & Grossman, J. B. (1998). Pragmatic directions and children’s word learning. Journal of Child Language, 25, 1–18.Find this resource:
Clark, E. V., & Hecht, B. F. (1982). Learning to coin agent and instrument nouns. Cognition, 12, 1–24.Find this resource:
Clark, E. V., & Hecht, B. F. (1983). Comprehension, production, and language acquisition. Annual Review of Psychology, 34, 325–349.Find this resource:
Clark, E. V., Hecht, B. F., & Mulford, R. C. (1986). Coining complex compounds in English: Affixes and word order in acquisition. Linguistics, 24, 7–30.Find this resource:
Clark, E. V., & Wong, A. D.-W. (2002). Pragmatic directions about language use: Words and word meanings. Language in Society, 31, 181–212.Find this resource:
Dromi, E. (1987). Early lexical development. Cambridge, U.K.: Cambridge University Press.Find this resource:
Dubois, J. (1962). Étude sur la dérivation suffixale en français moderne et contemporain. Paris: Librairie Larousse.Find this resource:
Duvignau, K., Fossard, M., Gaume, B., Pimenta, M. A., & Elie, J. (2007). Semantic approximations and flexibility in the dynamic construction and “deconstruction” of meaning. Lingugem em (Dis)curso, 7, 371–387.Find this resource:
Ellis, A. W., & Morrison, C. M. (1998). Real age-of-acquisition effects in lexical retrieval. Journal of Experimental Psychology: Learning, Memory, and Cognition, 24, 515–523.Find this resource:
Farrar, M. J. (1992). Negative evidence and grammatical morpheme acquisition. Developmental Psychology, 28, 90–98.Find this resource:
Fenson, L., Dale, P. S., Reznick, J. S., Bates, E., Thal, D. J., & Pethick, S. J. (1994). Variability in early communicative development. Monographs of the Society for Research in Child Development 59 [Serial No. 242].Find this resource:
Fox, B., & Routh, D. K. (1975). Analyzing spoken language into words, syllables, and phonemes: A developmental study. Journal of Psycholinguistic Research, 4, 331–342.Find this resource:
Gelman, S. A., Wilcox, S. A., & Clark, E. V. (1989). Conceptual and lexical hierarchies in young children. Cognitive Development, 4, 309–326.Find this resource:
Gentner, D., & Goldin-Meadow, S. (Eds.) (2003). Language in mind. Cambridge, MA: MIT.Find this resource:
Griffiths, P., & Atkinson, M. (1978). A “door” to verbs. In N. Waterson& C. Snow (Eds.), The development of communication (pp. 311–319). New York: Wiley.Find this resource:
Guilbert, L. (1975). La créativité lexicale. Paris: Librairie Larousse.Find this resource:
Hart, B., & Risley, T. R. (1995). Meaningful differences in the everyday experience of young American children. Baltimore, MD: H. Paul Brookes.Find this resource:
Hoff, E. (2006). How social contexts support and shape language development. Developmental Review, 26, 55–88.Find this resource:
Jusczyk, P. W. (1997). The discovery of spoken language. Cambridge, MA: MIT.Find this resource:
Kelly, B. F. (2011). A new look at redundancy in children’s gesture and word combinations. In I. Arnon & E. V. Clark (Eds.), Experience, variation, and generalization (pp. 75–89). Amsterdam: John Benjamins.Find this resource:
Kelly, B. F. (2014). The development of gesture, speech, and action as communicative strategies. Proceedings of the Berkeley Linguistics Society meeting (pp. 371–380). Berkeley, CA: Department of Linguistics.Find this resource:
Kempe, V., Brooks, P. J., Mironova, N. & Fedorova, O. (2003). Diminutivisation supports gender acquisition in Russian children. Journal of Child Language, 30, 471–485.Find this resource:
Lehrer, A., & Kittay, E. F. (Eds.) (1992). Frames, fields, and contrasts: New essays in semantic and lexical organization. Hillsdale, NJ: Lawrence Erlbaum.Find this resource:
Levelt, W. J. M. (1989). Speaking: From intention to articulation. Cambridge, MA: MIT.Find this resource:
Lyons, J. (1977). Semantics, vol. 1. Cambridge, U.K.: Cambridge University Press.Find this resource:
Marcos, H., Salazar Orvig, A., Bernicot, J., Guidetti, M., Hudelot, C., & Préneron, C. (2004). Apprendre à parler: Influence du mode de garde. Paris: L’Harmattan.Find this resource:
Morrison, C. M., Ellis, A. W., & Quinlan, P. T. (1992). Age of acquisition, not word frequency, affects object naming, not object recognition. Memory and Cognition, 20, 705–714.Find this resource:
Mulford, R. C. (1983). On the acquisition of derivational morphology in Icelandic: Learning about -ari. Íslenskt Mál, 5, 105–125.Find this resource:
Olson, J., & Masur, E. F. (2011). Infants’ gestures influence mothers’ provision of object, action and internal state labels. Journal of Child Language, 38, 1028–1054.Find this resource:
Olson, J., & Masur, E. F. (2013). Mothers respond differently to infants’ gestural vs. nongestural communicative bids. First Language, 33, 372–387.Find this resource:
Oviatt, S. (1980). The emerging ability to comprehend language: An experimental approach. Child Development, 51, 97–106.Find this resource:
Saxton, M., Backley, P., & Gallaway, C. (2005). Negative input for grammatical errors: effects after a lag of 12 weeks. Journal of Child Language, 32, 643–672.Find this resource:
Scollon, R. (1976). Conversations with a one year old. Honolulu: University of Hawaii Press.Find this resource:
Seva, N., Kempe, V., Brooks, P. J., Mironova, N., Pershukova, A., & Fedorova, O. (2007). Crosslinguistic evidence for the diminutive advantage: gender agreement in Russian and Serbian children. Journal of Child Language, 34, 111–131.Find this resource:
Slobin, D. I. (1978). A case study of early linguistic awareness. In A. Sinclair, R. Jarvella, & W. J. M. Levelt (Eds.), The child’s conception of language (pp. 45–54). Berlin: Springer-verlag.Find this resource:
Strapp, C., Bleakney, D. M., Helmick, A. L., & Tonkovich, H. M. (2008). Developmental differences in the effects of negative and positive evidence. First Language, 28, 35–53.Find this resource:
Vihman, M. (1996). Phonological development: The origins of language in the child. Oxford: Blackwell.Find this resource:
Vouloumanos, A., Martin, A., & Onishi, K. H. (2014). Do 6-month-olds understand that speech can communicate? Developmental Science, 17, 872–879.Find this resource:
Walker, D., Greenwood, C., Hart, B., & Carta, J. (1994). Prediction of school outcomes based on early language production and socioeconomic factors. Child Development, 65, 606–621.Find this resource:
Weisleder, A., & Fernald, A. (2013). Talking to children matters: Early language experience strengthens processing and builds vocabulary. Psychological Science, 24, 2143–2152.Find this resource:
Weizman, Z. O., & Snow, C. E. (2001). Lexical output as related to children’s vocabulary acquisition: Effects of sophisticated exposure and support for meaning. Developmental Psychology, 37, 265–279.Find this resource:
Winner, E. (1997). The point of words: Children’s understanding of metaphor and irony. Cambridge, MA: Harvard University Press.Find this resource:
Zevin, J. D., & Seidenberg, M. S. (2002). Age of acquisition effects in word reading and other tasks. Journal of Memory and Language, 47, 1–29.Find this resource: