Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, LINGUISTICS (linguistics.oxfordre.com). (c) Oxford University Press USA, 2016. All Rights Reserved. Personal use only; commercial use is strictly prohibited. Please see applicable Privacy Policy and Legal Notice (for details see Privacy Policy).

date: 17 December 2017

Phonological Templates in Development

Summary and Keywords

Child phonological templates are idiosyncratic word production patterns. They can be understood as deriving, through generalization of patterning, from the very first words of the child, which are typically close in form to their adult targets. Templates can generally be identified only some time after a child’s first 20–50 words have been produced but before the child has achieved an expressive lexicon of 200 words. The templates appear to serve as a kind of ‘holding strategy’, a way for children to produce more complex adult word forms while remaining within the limits imposed by the articulatory, planning, and memory limitations of the early word period. Templates have been identified in the early words of children acquiring a number of languages, although not all children give clear evidence of using them. Within a given language we see a range of different templatic patterns, but these are nevertheless broadly shaped by the prosodic characteristics of the adult language as well as by the idiosyncratic production preferences of a given child; it is thus possible to begin to outline a typology of child templates. However, the evidence base for most languages remains small, ranging from individual diary studies to rare longitudinal studies of as many as 30 children. Thus templates undeniably play a role in phonological development, but their extent of use or generality remains unclear, their timing for the children who show them is unpredictable, and their period of sway is typically brief—a matter of a few weeks or months at most. Finally, the formal status and relationship of child phonological templates to adult grammars has so far received relatively little attention, but the closest parallels may lie in active novel word formation and in the lexicalization of commonly occurring expressions, both of which draw, like child templates, on the mnemonic effects of repetition.

Keywords: accentual pattern, binomial, consonant harmony, distributional learning, phonological memory, phonological template, prosodic structure, rhythmic compound, typology

Child phonological templates are idiosyncratic word production patterns. They can be understood as deriving, through generalization of patterning, from the very first words of the child, which are typically close in form to their adult targets (Vihman, 2014). Thus templates can generally be identified only some time after a child’s first 20–50 words have been produced but before the child has achieved an expressive lexicon of 200 words (Renner & Strandberg, 2015); the interval between these lexical milestones typically constitutes a period of a few months but may last for as long as a year. The templates appear to serve as a kind of ‘holding strategy’, a way for children to produce more complex adult word forms while remaining within the limits imposed by the articulatory, planning, and memory limitations of the early word period (Vihman & Wauquier, in press). Templates have been identified in the early words of children learning a number of languages, although not all children give clear evidence of using them. Within a given language we see a range of different templatic patterns, but these are nevertheless broadly shaped by the prosodic characteristics of the adult language as well as by the idiosyncratic production preferences of a given child; it is thus possible to begin to outline a typology of child templates (Vihman, 2015, in press). However, the evidence base for most languages remains small, ranging from individual diary studies to rare longitudinal studies of as many as 30 children. Thus templates undeniably play a role in phonological development, but their extent of use or generality remains unclear, their timing for the children who show them is unpredictable, and their period of sway is typically brief—a matter of a few weeks or months at most. Finally, the formal status and possible relationship of child phonological templates to adult grammars has so far received little attention and remains uncertain. We will consider some possible parallels.

1 Developmental Profile and Time-Course

In the earliest period of word use a child must first remember adult word forms, along with their links to meaning, well enough to produce them in roughly appropriate situations; this makes it possible for these first words to be identified by caregivers (or investigators). The very first words are characteristically fairly ‘accurate’ (Ferguson & Farwell, 1975): they resemble their adult targets, which are themselves typically simple in structure—no more than two syllables long and generally featuring as syllable onsets glottals and glides, stops or nasals—and no more than a single consonant type per word form (e.g., bye, baby). The child forms largely match these targets with just the odd case of cluster reduction, omission of a coda, or substitution of a stop for a fricative. (See Appendix I, Menn & Vihman, 2011, for the first 5–6 recorded words of 48 children learning 10 languages.)

What are not observed at this stage are wholesale changes to the adult form, with rearrangement of segmental sequences (‘metathesis’) or truncation of syllables, for example. At some point after this, however, most children become more ‘ambitious’, attempting more challenging adult word forms that may be longer than the one or two syllables to which the early words are typically limited or that now include, in the adult form, a greater number of segments or sequences that the child is not yet producing, especially the challenge of consonant change across the form (e.g., banana, blanket, flowers, juice, kitty). This sets the stage for the emergence of phonological templates: having already deployed a number of recognizable words, the child now (implicitly, without conscious intent) generalizes the distributional patterning that underlies many of her words or—an alternative interpretation of the same observed behavior—falls back on the motoric routine needed for one or more of these early words to assimilate the more difficult target words that she is now attempting to produce. We discuss this mechanism below.

The more complex word targets that a child attempts at this stage are likely to be of high frequency in the input and also naturally salient due to their meaning or reference to objects or events that are inherently interesting to the child, their use in isolation, or their prosody; they are also typically similar, in some respects, to the child’s existing word forms, which provides a ‘hook’ for the imposition or projection of a templatic form. (All of these characteristics also favor the production of onomatopoeia, which commonly make up an important proportion of the early lexicon; see Laing, 2016.) In most cases more than one of these factors will play a role, leading the child to attempt production of a form despite the fact that in some respects it exceeds her articulatory, planning, or memory capacities. Each of these limiting factors is generally overcome only very gradually over the course of the single word period and thereafter, yet the rate of lexical learning may rise quite dramatically. For many children establishment of one or more whole-word phonological templates provides a launching pad for an increase in learning rate initially accompanied by a decrease in accuracy.

Thus phonological templates are the product of a particular moment in development, the window of time when a child’s experience of the world, of themselves as actors in the world, of the speech around them, and of their own vocal production combine to give rise to the frequently noted ‘vocabulary burst’ (Parladé & Iverson, 2011). In reality, although by no means not all children experience such a burst (Ganger & Brent, 2004), a sharp rise in the ability to retain and respond experimentally to minimal phonemic differences in novel word forms is generally seen by 17 months (Stager & Werker, 1997) while an equally impressive increase in words understood is seen a bit earlier, by 14 months (Bergelson & Swingley, 2012, 2013). This is roughly the period in which we first find phonological templates being deployed, although there is great variability in the timing, whether we compare children on the basis of chronological age or, more appropriately, size of expressive lexicon. In a diary study Vihman and Vihman (2011) report the application of a ‘palatal template’ to as much as 69% of one child’s words in the period of acquisition of 50 to 100 words, for example, with only a handful of words adapted to the pattern thereafter. In contrast, the children of Waterson (1971) and Priestly (1977) were each producing about 150 different word types by the time their phonological templates were observed.

2 Description and Illustration: The Classic Studies

The concept of child phonological templates plays no role, as such, in phonological theory derived from adult language. In fact, the idea of the child having any system at all based on his own production patterns rather than in terms of the system of generative rules relating adult target to child form was explicitly denied, for example, in Smith’s (1973) full-length monograph describing his son’s Acquisition of Phonology. The idea of child-specific patterning developed in the 1970s, the first period of intensive study of child phonology, although the term ‘template’ itself may not have been used before Menn’s (1983) overview of the field (see also Macken, 1995). Current formal approaches to phonological development, framed within Optimality Theory (OT), sometimes acknowledge the role of individual output patterns and the nonlinear path that is typically seen, from early accurate or ‘faithful’ word form production to the less accurate, ‘holistic’ patterning that the notion of template is meant to capture (e.g., Becker & Tessier, 2011, who model the induction of a harmony constraint in some detail for one child’s phonology). Others working within the currently popular OT tradition are entirely opposed to the notion of a template (e.g., Bat-El, 2010). However, selected illustrations of child rule, process, or constraint use, without quantification, are most often used to support this theoretical approach (e.g., Gnanadesikan, 2004). In contrast, within the ‘whole-word phonology’ tradition to which the notion of template as we are using it belongs, a full set of a child’s word types, if not tokens or variants—as produced in a given recorded session or, in a diary study, over a set period (e.g., a month or a period of acquisition of 50 or 100 forms)—is the preferred basis for analysis (see the papers collected in Vihman & Keren-Portnoy, 2013).

The concept of individual phonological templates arose from observation of early child words, as described in two landmark publications, Waterson (1971) and Menn (1971). Waterson’s description of the ‘nasal structure’, one of five ‘schemas’ she identifies for her son P, acquiring British English, is illustrated in Table 1 (the other schemas are the ‘Labial’, ‘Continuant’, ‘Sibilant,’ and ‘Stop’ structures). Critically, Waterson noted the difficulty of arriving at any satisfactory analysis of the child forms she had identified by the usual practice of aligning each child segment with its adult target and adducing ‘rules’ or ‘processes’ to relate the two (as later elaborated by Smith, 1973, following the model of Chomsky & Halle, 1968). Note, however, that Smith’s (1973) data collection began when his child was over 2 years old and just beyond the 200-word point in his expressive vocabulary, however, so that the theoretical disagreement between Waterson and Smith may, to some extent, have been based on an unequal comparison of two distinct developmental periods.

Table 1. P’s Phonological Template: <ɲVɲV> (British English, age 1;6)

Adult Target

Child Form

another

[ɲaɲa]

finger

[ɲe:ɲe:], [ɲɪ:ɲɪ]

Randall

[ɲaɲø]

window

[ɲe:ɲe:]

Source: Waterson (1971).

Menn’s (1971) case study of her son Daniel, acquiring American English, provides a similarly insightful analysis of a quite a different set of favored output patterns, although the study is couched in terms of a somewhat complex set of formal ‘phonotactic rules’. Menn begins her analysis with the very first words, identifying ‘Stage 1’ as the period from 16 to 22.5 months, which ends when Daniel had produced 29 words and the diarist parent was able to identify emergent patterning: ‘While phonotactic rules have not yet crystallized in stage 1, something vaguely systematic, from which the rules will develop, is at work’ (Menn, 1971, p. 232). Stage 2 (which she sees as ending at 24 months, when Daniel’s patterns change again) includes 53 words—nearly doubling Daniel’s vocabulary over a much shorter period.

From the ‘output skeleton’ Menn provides to indicate the ‘word types attested’ in stage 2 we see that Daniel had a clear preference for producing monosyllabic forms with codas (Table 2), which account for nearly 90% of his productions. If we treat as ‘selected’ any form that merely omits a segment not yet in repertoire, then only two processes are needed to describe Daniel’s adaptation of words to fit the template: truncation, for words of two or more syllables, and harmony, for words with a stop at both onset and coda, at different places of articulation. With the sole exception of banana [nænæ], all longer word forms are reduced to monosyllables in this period. The exhaustive listing of examples permits the reader to evaluate Menn’s hypotheses, which include the suggestion that ‘the nature of the phonotactic rules [or preferred output patterns] is partly determined by the shapes of the first handful of words attempted’ (1971, p. 246). Menn also concludes that ‘the word is an entity, stored and accessed as a block’ (p. 247)—a position that is well supported by the data that have since been reported within the ‘whole-word phonology’ approach.

Table 2. Daniel’s Phonological Templates: <CVC> (American English, age 1;10.15-2;0)

Adult Target

Child Form: selected

Adult Target

Child Form: adapted

Processes applied

ant

[ænt]

apple

[æp]

truncation

bump

[bʌmp]

berries

[bejz]

truncation

car

[gar]

boat

[bop]

harmony

change

[ejndʒ]

bread

[bʌb]

harmony

corn

[gɔ(r)n]

bug

[gʌg]

harmony

eat

[it]

cracker

[gæk]

truncation

horse

[ars]

drink

[gɨnk]

harmony

hug

[ʌg]

glasses

[gæs]

truncation

mouse

[mæws]

meat

[dit]

harmony

move

[muf]

moon

[mum]

harmony

noise

[nɔjz]

napkin

[næk]

truncation

night

[najt]

pacifier

[bæf]

truncation

rake

[ejk]

radish

[æs]

truncation

ride

[ajd]

sand

[æn]

shoes

[uz]

Steve

[iv]

stone

[don]

swing

[iŋ]

toast

[dos]

watch

[aʃ]

Source: Menn (1971).

Priestly (1977) provides what remains one of the most complete and persuasive illustrations of a child phonological template (see Table 3). The study follows one child’s phonological and lexical advances from the onset of noticeable systematicity in production, eventually resulting in 70 ‘bisyllabic experimental forms’, to their gradual replacement, over four months of use, by ‘ordinary replacement forms’ that featured the expected cluster reduction and other early substitutions for difficult segments and sequences but not the idiosyncratic pattern of the intervening period. (In two appendices Priestly provides both early and later forms in full.)

Table 3. Christopher’s Phonological Template: <CVjVC> (British English, age 1;10-2;2)

Adult Target

Child Form

berries

/ˈbɛrijz/

[bɛjas]

chocolate

/ˈtʃɒklɪt/

[kajak]

lion

/ˈlɑjən/

[lajən]

peanut

/ˈpijnʌt/

[pijat]

tiger

/ˈtajgə/

[tajak]

whale

/ˈwɛjəl/

[wɛjəl]

Source: Priestly (1977).

The last of the ‘classic studies’ of the template phenomenon is Macken’s (1979) account of the acquisition of Mexican Spanish by a girl she calls Si. This longitudinal observational study traced the child’s deployment of a ‘melody’, or fixed segmental sequence—labials followed by coronals—over a six-month period. For example, at 1;8 the child produced her older brother’s name, Fernando, as [ˈman:ə], blending the labiality of the word-initial consonant with the nasality of the stressed syllable onset; at the same age she produced sopa ‘soup’ as [ˈpʻwætʻa], with metathesis to achieve the preferred sequence, labial–coronal. At 1;10 she produced vestido ‘dress’ as [ˈbɪt:ɪ], shortening the target to fit the disyllabic template, maintaining the vowel of the second (stressed) syllable but retaining the labiality of the onset. Macken meticulously describes the ways in which the pattern was enlarged, changed, and ultimately abandoned as the child’s ability to remember, plan, and produce more complex sequences grew along with her lexical advances.

It is notable that the patterns described—the first whole-word templates to have been identified—are quite different in these four children. Waterson’s son P displayed small word groups (or ‘schemas’; Waterson, 1971) with distinct templatic patterns, some monosyllabic (e.g., the ‘sibilant structure’ illustrated below), others disyllabic (the ‘nasal structure’; Table 1), each featuring a single preferred consonantal manner. Menn’s son Daniel displayed a strong preference for closed monosyllables, while Christopher Priestly, at a more advanced lexical level, deployed up to two consonants in his disyllabic word forms with codas but avoided the challenge of producing the full set of target consonants (e.g., in chocolate). Finally, Macken’s subject Si tackled the typically multisyllabic forms of her language, Spanish, which are often longer than two syllables, by developing a fixed consonantal sequence or melodic template and truncating targets to arrive at a disyllabic form in most cases.

3 Criteria for Template Identification and Further Cross-Linguistic Illustration

Templates constitute a subset of a child’s most often produced prosodic structures but stand out from those structures in two ways: First, they may be identifiable by the child’s overuse (or overselection) of certain patterns in comparison with other children learning the same language, or by their adaptation of target forms to fit the constraints of the preferred pattern. Vihman (2016) provides a detailed account of template identification and use in relation to prosodic structures, based on diary studies of five bilingual children; Vihman (in press) shows the effects of both ‘selection’ and ‘adaptation’ in the prosodic structures themselves as deployed by 44 children learning four languages.

Second, whereas prosodic structures are identified on the basis of their sequencing of consonants and vowels alone—i.e., on the basis of their C-V skeleton—templates may be further specified segmentally. That is, they may feature particular segments or segment classes in particular word positions (e.g., medial yod; Table 3) or sequences (e.g., nasal harmony; Table 1; labial or velar harmony; Table 2; melody, Macken’s Si).

Vowel choices are seldom involved in child templates. This may be because vowels are well practiced earlier in development than consonants, from before the emergence of canonical babbling. Vowel production may nevertheless be highly variable; achievement of an exact match does not appear to have high priority for either the child or the caretakers, and indeed some children (e.g., Joan, in Velten, 1943; Timmy, in Vihman, Velleman, & McCune, 1994) may include in their word forms only the low unrounded vowel [a] or [ɑ] of the first canonical babbling for several months of vocabulary building. (The effect, in canonical babbling as produced in an infant’s small oral cavity with a proportionately large tongue, is of a somewhat fronted vowel. Note that infant vowels are also notoriously difficult to transcribe reliably; some relevant phenomena may simply be missed.)

Vowel harmony has rarely, if ever, been reported as a child template, and vowel melodies, a parallel to the consonantal melody described for Macken’s Si, are also infrequently observed (but see Vihman, 1976, for a <low V . . . high V> template). Two examples of templates primarily affecting the vocalic nucleus both involve the same palatal template <CVi> (Alice, acquiring American English, in Vihman et al., 1994; Maarja, acquiring both Estonian and English, in Vihman & Vihman, 2011).

The use of templates generally corresponds to a ‘regression in accuracy’ that can be taken as an indicator of phonological reorganization or systematization. (Within an OT account, this amounts to unexpected demoting of ‘faithfulness’: see, for example, Becker & Tessier, 2011; Bernhardt & Stemberger, 1998; Fikkert & Levelt, 2008) The earliest period of word use is dominated by ‘selection’, or production that attempts only the most accessible targets. Accordingly, the first words tend to be more or less accurate renditions of those targets, within the limits of the child’s resources. However, once a child has extracted a favored pattern or template from the words he is regularly producing or developed one or more stable, readily deployed articulatory or motoric routines, he often begins to produce more difficult words in a far less accurate form (see Daniel’s adapted forms, Table 2). Since memory for sequences is likely to be an issue as much as articulatory complexity, this developmental shift from ‘selecting’ to ‘adapting’—which then occurs in parallel with additional ‘selecting’—can be understood as reflecting the increasing power of a child filter that renders particularly salient and memorable those aspects of input speech forms that map more or less faithfully onto his own preferred production patterns.

A consequence of this shift from ‘select’ to both ‘select’ and ‘adapt’ is that a child’s forms begin more and more to resemble one another under the influence of the favored output pattern; this can lead to considerable homonymy, as noted by Velten (1943), who saw this as a consequence of the slow ‘growth of the phonemic pattern’ alongside rapid vocabulary expansion (p. 83). This does not typically pose a problem at this developmental level, when the children are most often speaking of the here-and-now to interlocutors familiar with their interests and preferences. (See Vihman [1981, 2014, App. 3, and 2016] on the ‘homonym strategy’ observable in the words produced in the early period by children like Waterson’s son P, Hildegard Leopold, and Vihman’s son Raivo, all of whose small templatic word groups resulted in a good deal of homonymy alongside increased diversity in the word types attempted.)

Cross-linguistically, many of the same templatic patterns or ‘solutions’ to phonological challenges can be seen (Vihman, in press). For example, templates developed early on often rely on simple monosyllabic forms, like Daniel’s <CVC> and Maarja’s <CVV>. Vihman and Croft (2007) cite diary studies to illustrate monosyllabic templates that require a final sibilant. These are <(Stop)Vʃ>, P’s ‘sibilant structure’ at 1;6 as described by Waterson (1971); brush [byʃ], fetch, fish [ɪʃ], vest [ʊʃ] and <(Stop)Vs:>, observed in the first 50 words of Madli, 1;8, acquiring Estonian; isa, issi ‘daddy’ [is:], suss ‘slipper,’ and uss ‘snake’ [us:]. Vihman’s Estonian- and English-learning son Raivo displayed, at 1;3, a similar template <CoVFric>, more broadly specifying ‘fricative’ in the final slot, which allowed not only for words like küpsis ‘cookie’ [kys], musi ‘kiss,’ and müts ‘hat’, both [məs], this [dɪs], and juice [zøs], [jøs], [ʒus], but also juust ‘cheese’ [duɬ’], [uf] (Vihman, 1981, 2014, 2016).

However, the first templatic pattern of many children is simply <CV> (see Camille, learning Swiss French, Table 4). Notice that here again some forms are accurate, based on simple adult open-monosyllable targets, more common in French than in English, while others involve truncation (attends, bébé, canard, etc.). In yet other cases, however, the child has ‘found’ within the target word a CV syllable that she can articulate only by radically adapting the target, imposing her preferred template as a selective filter on the adult sequence: e.g, gateau, where the onset velar is blended, in the child form, with the medial coronal, and musique, where the final syllable is metathesized.

Table 4. Camille’s Phonological Template: <CV> (Swiss French, age 1;6)

Adult Target

Child Form: Selected

Adult Target

Child Form: Adapted

chat ‘cat’

/ʃa/

[ʃa]

attends ‘wait!’

/atã/

[tã]

chaud ‘hot’

/ʃo/

[(t)ʃo]

bébé ‘baby’

/bebe/

[be]

cloun ‘clown’

/klun/

[ku]

canard ‘duck’

/kanaʁ/

[ka]

main ‘hand’

/mɛ̃/

[mɛ̃]

chercher ‘look for’

/ʃeʁʃe/

[ʃe]

non ‘no’

/nõ/

[nõ]

cheveux ‘hair’

/ʃ[ə]vø/

[ʃø]

pas ‘not’

/pa/

[pa]

couvercle ‘lid’

/kuvɛʁkl/

[ku]

peur ‘(I’m) scared’

/pøʁ/

[pø]

encore ‘again, more’

/ãkoʁ/

[ko(ʁ)]

pied ‘foot’

/pje/

[pe]

enlever ‘to take out’

/ãl(ə)ve/

[ve]

plus ‘[no] more’

/ply/

[py]

fermé ‘closed’

/feʁme/

[fe]

tiens ‘take (it)’

/tjɛ̃/

[ta]

gâteau ‘cake’

/gato/

[ko]

musique ‘music’

/myzik/

[ki]

tombé ‘fall’

/tõbe/

[be]

Similarly, broad disyllabic patterns such as reduplication are also observed, as in the word forms of a precocious child, Annalena, acquiring German (Table 5).

Table 5. Annalena’s Phonological Template: <σ‎‎1σ‎‎1> (German, age 0;10-1;0)

Adult Target

Child Form: Selected

Adult Target

Child Form: Adapted

pipi ‘peepee’

[pipi:]

Annalena

[nana]

kikeriki ‘cock-a-doodle-do’

[ki:ki:]

wauwau ‘bowwow’

[vava]

tag ‘(good)day’

[dada]

zahn(bürste) ‘tooth(brush)’

[nana]

Sources: Elsen (1996) and Vihman and Croft (2007).

The most common template observed cross-linguistically is consonant harmony, which constitutes a simple solution to the challenge of producing (remembering, planning, and articulating) more than one consonant over a one- or, more often, two-syllable stretch under a single prosodic contour. Menn (1971) was the first to detail the systematic use of harmony, in the diary study described above. (See also Stoel-Gammon & Cooper, 1984, who report three case studies, one of them another American English–learning Daniel, who showed strong velar harmony in monosyllables.)

Table 6 provides an example of harmony in disyllabic forms from Finnish, whose relatively restricted consonant inventory seems to lend itself particularly readily to this templatic pattern (Vihman, 2015, in press).

Table 6. Eliisa’s Phonological Template: <C1VC1V> (age 1;5)

Adult Target

Child Form: Selected

Adult Target

Child Form: Adapted

kukka ‘flower’

[kuk:a]

lintu ‘bird’

[titu]

pupo ‘rabbit’

[pupo]

nuke ‘doll’

[kuk:e]

pallo ‘ball’

[pap:u]

rikki ‘broken’

[kik:i]

vettä ‘water’

[tit:æ]

Sources: Kunnari (2000), Vihman and Kunnari (2006), and Vihman & Velleman (2000).

Consonant harmony, which occurs only rarely in adult languages (as compared with vowel harmony, for example; Hansson, 2010, p. 19), has been the topic of a good deal of discussion in the formal phonological literature, in relation to both adults (e.g., Hansson, 2010; Mackenzie, 2009; Shaw, 1991) and children (Berg, 1992; Bernhardt & Stemberger, 1998; Goad, 1997; McAllister Byun, & Inkelas, 2014; Pater & Werle, 2003; Stoel-Gammon & Stemberger, 1994; Vihman, 1978). Smith’s son Amahl continued to make heavy use of the pattern at an age and vocabulary level when templates have typically faded; Smith was sufficiently impressed by the phenomenon to identify it as a ‘universal’ of child phonology. In fact, it is of very widespread occurrence in early child production (see also, for example, the Spanish child Jo in Macken, 1978), but some children fail to avail themselves of it at all. Like most aspects of child phonology, usage is variable, with individual differences by child playing a role as well as aspects of the adult language, especially prosodic and accentual factors.

4 Typology of Templates

The phenomenon of template use has now been identified in a wide range of languages, including Arabic (Khattab & Al-Tamimi, 2013), American English (Vihman & Velleman, 1989), Catalan (Lleó, 1990), Finnish (Savinainen-Makkonen, 2007; Vihman & Velleman, 2000), French (Brulard & Carr, 2003; Vihman, 1993; Wauquier & Yamaguchi, 2013), German (Kehoe, 2015), Hebrew (Keren-Portnoy & Segal, in press), Hindi-Urdu (Bhaya Nair, 1991; see Anwaar, 2015; Vihman & Croft, 2007), Italian (Keren-Portnoy, Majorano, & Vihman, 2009; Vihman & Majorano, in press), Pashto (Anwaar, 2015), Polish (Szreder, 2013), Brazilian Portuguese (Baia, 2013; Oliveira-Guimarães, 2013), and Castilian Spanish (Kehoe, 2015; Lleó, 1990), in addition to the studies already cited for American and British English, Estonian, French, German, and Mexican Spanish. Most of these more detailed accounts are case studies of a single child or at most 3–5 children; Vihman et al. (2013) report templatic patterning in 32 children, 21 of them identified at age 2 as ‘late talkers’. Despite the limited sampling, however, we can begin to identify differences in the templates of children learning different languages.

The forms children produce in the early word period are similar cross-linguistically, since the developmental limitations on motoric skill, articulatory planning, and phonological memory are all rooted in neurophysiological immaturity and infants’ initial lack of vocal experience. Each individual child overcomes these limitations in his or her way, presumably on the basis of differences in inherent verbal skills or interest in language, social scaffolding, and other unknown, perhaps unknowable factors. However, the prosodic structures and accentual patterns of the ambient language do affect children’s overall word shapes, resulting in identifiable typological constraints on templates by language (Vihman, 2015, in press). We will cite three well-established examples.

4.1 English Templatic Pattern: Disyllables with Codas

The templatic pattern of disyllables with codas, illustrated in Table 3, is unknown in the child Romance-language or Western Finnic data reported to date but is not uncommon in English. For example, of the 21 children whose late start on word production was reported in Vihman et al. (2013), at least 5 showed adaptation of words to fit a <CVCVC> template (Table 7). Because these children were older than is typical for this developmental milestone and were slow only in expressive vocabulary acquisition, not in comprehension, their templates tend to be more complex than those of the younger children whose early words are shown above. Note that every one of the child templates illustrated includes consonant harmony across two of the three or more consonants in this structure, and also that the coda is segmentally specified for two of the children. All of these specifications give further indications of the challenge such a structure poses; each of them further reduces the options in ways that increase efficiency in all aspects of production, including accessing the memory trace and planning and articulating the form.

Table 7. Disyllabic Phonological Templates with Word-Final Coda, British English

Adult Target

Child Form (selected)

Adult Target

Child Form (adapted)

a. Clarissa, 2;3, <C1VC1VC>

baby

[bebɪʔk]

bananas

[mɑ:mak]

chicken

[geɪgɪp]

icecream

[gɪgɪm]

jelly

[lɛlɪp]

b. Dean, 2;4, <C1VC1Vs>

saucers

[tatas]

ambulance (I)

[bæbɪs]

baby (I)

[beɪbɪs]

flowers

[fɑ:væs]

lady (i)

[dɪ:dɪs]

more animals(i)

[mɔmɔ:s]

c. Simon, 2;5, <CVCVn>

sit down

[a:didaʊn]

again

[dəden]

that’s it (I)

[dədɪt]

Dillon

[dɜ:nɪnɪn]

Notes: (*) The session from which these data are taken was in each case the first in which the child used at least 25 word types spontaneously.

Source: Vihman et al. (2013).

4.2 ‘Null Onset’ Pattern in Languages with Iambic Accent

The two remaining examples of adult language shaping of child templates both involve the pattern <VC(C)V>, which occurs in iambic-accent languages and in languages with geminates but which has not so far been observed in the far larger data set available for children learning English. French, for example, consistently accents phrase-final syllables, rendering that syllable prosodically salient when a word is produced in isolation or in phrase-final position (see Wauquier & Yamaguchi, 2013, who argue strongly that the accent in French is not truly ‘iambic’ because French lacks word-level stress). Children learning French frequently use the pattern (see Charles in Vihman & Kunnari, 2006, for example, whose pattern is <aCa/o>). The most striking data come from Beryl, a child who produced 32 out of 56 words (57%), fitting her template in a half-hour recording session (Wauquier & Yamaguchi, 2013). Table 8 provides several examples, including many showing metathesis (e.g., baleine, crapeau, nuage, parrain, sardine).

Table 8. Beryl’s phonological template: <aCV> (French, age 1;8)

Adult Target

Child Form: Selected

Adult Target

Child Form: Adapted

agneau ‘lamb’

/aɲo/

[alo]

(l)éléphant

/elefɑ̃/

[afɔ]

encore ‘again’

/ɑ̃kɔʁ/

[ɑ̃kɔ]

baleine ‘whale’

/balɛn/

[anɛ]

hibou ‘owl’

/ibu/

[abu]

cerceau ‘circle’

/sɛʁso/

[aço]

crapeau ‘toad’

/kʁapo/

[ako]

étoile ‘star’

/etwal/

[ata]

les cloches ‘the bells’

/leklɔʃ/

[aχ‎‎lo]

musique ‘music’

/myzik/

[aɬi]

nuage ‘cloud’

/nɥaʒ/

[aça]

parrain ‘godfather’

/paʁɛ̃/

[apa]

sardine ‘sardine’

/saʁdin/

[ani]

Source: Wauquier and Yamaguchi (2013) and Vihman and Wauquier (in press, Appendix 1a).

A <VCV> template also occurs in the word forms of child learners of both Hebrew (Keren-Portnoy & Segal, in press) and Pashto (Anwaar, 2015), both of which have a predominantly iambic pattern in the more usual sense of word stress on the second syllable of disyllabic forms.

4.3 ‘Null Onset’ Pattern in Languages with Medial Geminate Clusters

A <VCV> pattern is also observed in languages with geminates, including Hindi-Urdu (Anwaar, 2015; Vihman & Croft, 2007) and Estonian (Vihman, 2016).1 Tables 9 and 10 illustrate the pattern from children learning Finnish and Italian. Note that Italian children sometimes use truncation to obtain the preferred disyllabic word length, since in Italian, as in Spanish, the basic vocabulary includes many longer words than are found in Dutch, English, French, German, or Swedish, for example.

Table 9. Atte’s phonological template: <VC(C)V> (Finnish, age 1;8: 16/26 words[62%] fit template)

Adult Target

Child Form: Selected

Adult Target

Child Form: Adapted

ääni ‘sound’

/æ:ni/

[æ:ni]

kala ‘fish’

/kala/

[ala]

äiti ‘mother’

/æiti/

[æit:i]

kello ‘clock’

/kɛl:o/

[el:o]

ankka ‘duck’

/aŋk:a/

[ak:a]

loppu ‘all gone, finished’

/lop:u/

[op:u]

auto ‘car’

/auto/

[auto]

nalle ‘teddy’

/nal:e/

[al:e]

isi ‘daddy’

/isi/

[içi]

pallo ‘ball’

/pal:o/

[al:o]

Note: (*) Stress is consistently on the first syllable in Finnish.

Source: Kunnari (2000) and Vihman and Wauquier (in press, Appendix 1b).

Table 10. J.O.’s phonological template: <VCV> (Italian, age 1;9: 23/69 words [33%] fit template)

Adult Target

Child Form: Selected

Adult Target

Child Form: Adapted

acqua ‘water’

/ak:wa/

[ak:wa]

bella ‘pretty, nice, f.’

/ˈbɛl:a/

[ɛl:a]

altro ‘other, m.’

/altro/

[at:o]

coltello ‘knife’

/kolˈtel:o/

[el:o]

apri ‘open!’

/apri/

[api]

forchetta ‘fork’

/forˈket:a/

[et:a]

ecco ‘here (it is)!’

/ɛk:o/

[ɛk:o]

freddo ‘cold, m.’

/ˈfred:o/

[ed:o]

giallo ‘yellow’

/ˈdʒal:o/

[al:o]

maialino ‘piggy’

/maiaˈlino/

[ino]

pecorella ‘lamb’

/pekoˈrel:a/

[el.a]

polpetta ‘meatball’

/polˈpet:a/

[et:a]

pronto ‘ready, m.’

/ˈpronto/

[onto]

questo ‘this, m.’

/ˈkwesto/

[et:o]

Source: Vihman and Majorano (in press).

5 Origins of Templates: ‘Secondary Distributional Learning’

We have considered examples of templatic patterning in children learning several languages and noted that the templates tend not to be observed in the very first words but emerge at some point after the child has made a solid start on regular word production. What is the mechanism behind template formation? Any answer will necessarily be speculative; we lack relevant experimental or electrophysiological evidence, and it is in any case unclear how either approach could shed light on how the children arrive at these patterns. Modeling the extraction of templatic patterns from child word forms can be informative but has not often attempted (exceptions include Becker & Tessier, 2011; Berg & Schade, 2000; Menn, Schmidt, & Nicholas, 2013, discuss what must go into such a model). Furthermore, template formation, with the regression in accuracy that it entails, has the effect of children distancing themselves from the adult language that they can reasonably be thought to be attempting to learn (and accordingly to some degree foiling their own presumed communicative goals) just as word learning is becoming well established. The best way to understand this paradox, we argue, is to see it as an unconscious manifestation of the inescapable human (and more broadly animal) tendency to seek out and generalize from patterns or regularity in every aspect of experience.

The role of distributional (‘statistical’ or probabilistic) learning, based on implicit sensitivity to recurrent sequences and patterns, has received increasing attention in recent years, particularly since Saffran, Aslin, and Newport (1996), a study that dramatically shifted our understanding of what ‘learning’ might mean (see also Kelly & Martin, 1994, who anticipated the findings). Saffran et al. (1996) demonstrated that after a mere two minutes of ‘training’, or auditory exposure to an uninterrupted, monotone repeating string of syllables in which four ‘words’, triplets of set syllabic sequences, were embedded, 8-month-olds responded differentially at test to ‘grammatical’ vs. ‘ungrammatical’ triplets—i.e., to nonwords that matched the four that constituted the training stimulus string from nonwords made up of the same syllables but in sequences that did not recur in training and so did not lend themselves to extrapolation as a ‘word’ or unit. Specifically, the infants showed a novelty response, attending longer at test to the nonwords that had not been experienced as likely units in the training string. A similar effect, demonstrating infant sensitivity to co-occurrence probabilities, was also shown in the visual modality (Kirkham, Slemmer, & Johnson, 2002). And the same original experiment, using a different procedure but with nonword strings similarly created and presented as auditory stimuli, also successfully demonstrated ‘incidental learning’ in older children and in adults, who remained unaware of the patterning until tested on it (Saffran, Newport, Aslin, Tunick, & Barrueco, 1997).

Since then, numerous studies have been conducted to explore the nature and extent of such implicit, incidental, unconscious learning—in other words, learning that occurs outside of focused attention. One striking study (Maye, Werker, & Gerken, 2002) demonstrated that infants’ ability to discriminate phonetic categories could be shaped in the course of a brief experiment that exposed them to syllables synthesized to evenly represent a distribution along the acoustic continuum from (prevoiced)/da/to (voiceless unaspirated)/ta/. The stimuli were presented to two groups of infants, bimodally (with more exemplars taken from the extremes of the continuum than from the center) to one group, unimodally (exemplars clustering in the center of the continuum) to the other. The infants who had experienced bimodal exposure gave evidence of discriminating the syllables at test; the infants with unimodal exposure failed to show discrimination.

Some researchers argue in favor of a major role in infant phonological development for distributional learning—not only of sequences, but also phonological categories, formed from the acoustic clustering thought to reflect phonemic contrasts (e.g., Kuhl et al., 2008). Others urge caution in the interpretation of these somewhat artificial experimental findings and question whether distributional learning can ‘scale up’ from studies of a few test stimuli presented under tightly controlled conditions (e.g., Johnson, 2012; Lin & Mielke, 2008; McMurray & Hollich, 2009; Swingley, 2009).

There is little doubt that distributional learning plays a role in language development, familiarizing the child with the sounds and sequences of ambient speech. However, the link between implicit learning of recurrent sequential patterns or the clustering of sounds into contrasting categories, on the one hand, and the knowledge of whole-word forms needed for production, on the other, remains unclear, for several reasons. First of all, learning to produce identifiable word forms under the appropriate situational conditions requires, especially in the early stages of referential word use (McCune & Vihman, 2001), that the child grasp the arbitrary association of word forms with their meanings. And words must be accessible not only in specific ‘priming’ contexts, which often support first word use, but more broadly, across situations of use, with certain word exemplars being understood as counting as members of unitary categories of meaning that serve for reference (i.e., different word token occurrences or variants may belong to a single word type).

Declarative learning is widely viewed as a distinct learning mechanism from the neocortical, sensorimotor processing that, with repeated exposure, gradually results in implicit learning, including ‘statistical’ or ‘distributional’ learning (for more on the complementary systems account of word learning, see, e.g., Huber & Born, 2013; Lindsey & Gaskell, 2010; McClelland, McNaughton, & O’Reilly, 1995). In adults, the learning of arbitrary form-meaning matches involves ‘episodic memory’, the product of both the hippocampus, which rapidly binds disparate elements of an experienced moment in time (critical for associating spoken word forms with aspects of a situational experience), and the prefrontal lobes, which guide attention. In children in the early stages of language acquisition, it has been proposed that fast mapping—identification of an unknown word form association with a referent in the context of a contrasting known word with a related meaning (Carey, 1978)—may enable rapid word learning through neocortical structures (Merhav, Karni, & Gilboa, 2015; Sharon, Moscovitch, & Gilboa, 2011) in the absence of a fully developed hippocampal system (Bauer, 2008). In either case the child’s attention must be focused on the form-meaning relationship for long-term declarative memory, or specific item learning, to ensue. Thus, before producing words referentially, in more than one potentially appropriate type of context, children’s word form learning must derive to a great extent from words heard in situations of focused attention, when their own actions and those of others can leave memory traces in which referents are associated with co-occurring speech forms.

More to the point here, the kinds of word forms we have reviewed strongly suggest that infants do not learn sounds and then combine them into words, as some models would imply; instead, they learn whole-word patterns and only gradually develop a network based on connections between phonologically related words of various kinds (same onset, same rhyme, etc.; Vihman, 1981, 2016; Werker & Curtin, 2005, draw similar conclusions based on evidence from perception alone). Distributional learning can provide only a small part of the knowledge needed for building a lexicon; it is hard to see how it would provide the word form knowledge required for production.

Furthermore, despite the large literature on infant ability to segment words from running speech (Jusczyk, 1997), there is reason to believe that words heard in isolation, while admittedly constituting only a fraction of the speech infants hear, are disproportionately important, especially for the earliest stages of word learning (Bortfeld, Morgan, Golinkoff, & Rathbun, 2005; Brent & Siskind, 2001; Lew-Williams, Pelucchi, & Saffran, 2011). For example, of the first words of the 17 English-learning children listed in the Appendix to Menn and Vihman (2011), fully one-third are ‘non-syntactic’ forms such as hello, thanks, uh-oh, and various onomatopoeia; every child’s first-word set includes at least one such word. These words could not be learned through segmentation but only from repeatedly hearing them in isolation in the course of everyday routines.

It has been argued, finally, that the very first words are likely based on infant matching of their own vocal forms to often repeated words in the input (Vihman, 1993); recent studies have provided plausible experimental evidence to support that hypothesis (DePaolis, Vihman, & Keren-Portnoy, 2011; DePaolis, Vihman, & Nakai, 2013; Majorano, Vihman, & DePaolis, 2014; see also Vihman, DePaolis, & Keren-Portnoy, 2014). This then would provide, at least for relatively early talkers whose receptive vocabulary may be only a little ahead of their expressive vocabulary, the very first words that a parent or investigator can observe or record.

Once the child has begun using a few words with some regularity, a new internal ‘database’ is automatically created, based on the representations of these words. The same distributional learning mechanism that operates on input speech and other experiences could be expected to operate on this database as well, and to do so all the more effectively because the child’s own word production is likely to be better retained in memory, given the additional effort of production and the effects of repeated practice, than the fleeting experience of others’ speech (Elbers & Wijnen, 1992; Gershkoff-Stowe, 2002; Keren-Portnoy, Vihman, DePaolis, Whitaker, & Williams, 2010; see also recent experimental studies of ‘the production effect in memory’ in both adults and older children, e.g., Icht & Mama, 2015; MacLeod et al., 2010; Zamuner, Morin-Lessard, Strahm, & Page, 2016). We argue, then, that ‘secondary distributional learning’ is the most likely mechanism for the emergence of templates: this is systematization arrived at as a function of the child’s (temporary) shift to more heavily weighting structures that integrate their own vocal patterns with the target forms presented by the ambient language. (An alternative interpretation sees this shift in weighting in terms of competition with ongoing form learning from a ‘cache’ of previously used word forms; Becker & Tessier, 2011.)

6 Relation of Child Phonological Templates to Linguistic Theory

Templates are ‘output-oriented’, in the sense that a variety of different phonological processes appear to conspire to arrive at a single favored child pattern (see Daniel Menn’s complementary application of truncation and harmony; Table 2). As such they should in principle fit in with many current phonological models, particularly OT (Tesar, 2013). However, in her full-length introduction to both Phonological Acquisition and OT, Tessier (2016) briefly mentions child phonological templates but illustrates the concept with a description of Priestly’s data only; she does not discuss templates more broadly as relating to a plausible approach to understanding phonological development.

The evidence that whole-word patterns play a key role in phonological development and that they emerge only after a period of individual item learning is in better accord with usage-based exemplar models of adult grammar (e.g., Barlow & Kemmer, 2000; Bybee, 2001, 2006; Bybee & Hopper, 2001; Bybee & Slobin, 1982; Ellis, 2002, 2005; Pierrehumbert, 2003; see also Vihman & Croft, 2007). These models share the overall perspective of the distributional learning models but derive from linguistic analysis as much as from computational modeling and are oriented more often toward diachronic than toward developmental accounts. The emphasis of these models on exemplars in memory fits well with the evidence of emergent and variable patterning that we see in individual child templates. Furthermore, their focus on ‘performance’ or usage and its continuous transformation, through repeated experience, into ‘competence’ or knowledge provides a good theoretical framework for understanding the phenomenon at issue here. This is in contrast with the sharp distinction that continues to be drawn between competence and performance in generative approaches (see Tessier, 2016).

Reflecting on their experimental work with older children, Munson, Edwards, and Beckman (2012) describe a process of generalization based on initial item-learning, or instance-based exemplars, that is much like the ‘secondary distributional learning’ we have described. They have established that lexical learning plays an important role in nonword repetition, a strong predictor of lexical advance (Gathercole & Baddeley, 1993), which they ascribe to the emergence of phonological representations that support access to phonotactic mappings for novel (native-language) lexical forms (see also Vihman, 2014, ch. 11; Vihman et al., 2014).

7 Relation of Child Phonological Templates to Adult Grammar: Non-Concatenative Morphology and Reduplication with Fixed Segmentism

What is less evident is the kind of grammatical entity that we might turn to for an adult parallel to child templates. Clearly the much-discussed root-and-pattern structure of Semitic languages—Hebrew and Arabic, most prominently—are related in a formal but somewhat distant sense to the notion of template as used here. McCarthy (1981) initiated the first formal work on non-concatenative structure in Arabic. The idea of a CV template, ‘the sekeletal structure of the word that abstracts away from the segmental features of phonemes, retaining only their status as consonants or vowels’ (Dawdy, Hesterberg, & Pierrehumbert, 2014, p. 1269), grew out of that work; it does not exactly correspond to the use of the term ‘template’ in Semitic studies, where it refers to the ‘whatever is left of the word after the root is extracted’ (Shimron, 2003, p. 9), not to the skeleton as a whole (<CVCVC>, for example). The child templates that we have described are specifiable mainly in terms of the C-V skeleton, but specification of particular segments is one of their identifiable characteristics as individual children adapt words to fit their pattern. Needless to say, in addition to the formal differences in the way the term is defined, the uses of Semitic templatic patterning are quite different from child uses because they serve to make grammatical distinctions of word class, number, tense, and so on, whereas the child forms are purely lexical.

Subsequent work has sometimes extended the notion of templates to aspects of languages with concatenative morphology. For example, Scheer (2003) takes the Semitic template as a model for analysis of the apparently idiosyncratic Czech vowel length; Carvalho (2004) uses templatic morphology in an analysis of the Portuguese verb. However, the templates identified in these studies of languages outside the Semitic family become evident only as the result of linguistic ‘excavation’, digging below the surface to find patterns well hidden from the contemporary user. Templates identified in this way appear even less similar to child templates, in that they scarcely seem to play a productive role in language use that could be taken to parallel what we observe in the case of children’s templates, although at earlier (diachronic) stages of a given language, when the process must have been more transparent, what is now a deeply grammaticalized template may have had a more prominent ‘on-line’ cognitive function.

Another parallel to child templates is what is known as ‘reduplication with fixed segmentism’ (Inkelas & Zoll, 2005). Here again the template-like patterning has grammatical or meaning-changing function, as in this example from Goswami (1955–1956, p. 164, as cited in Alderete et al., 1999, p. 328): gharā ‘horse’, ghara-sarā ‘horse and the like’; khori ‘fuel’, khori-sori ‘fuel and the like’. Reduplication of this kind is also known as ‘melodic overwriting’; the fixed element is part of an affix that copies all or part of the base to which it attaches but ‘overwrites’ the corresponding segment in the base with that fixed element, as in the example given. Interestingly, Inkelas and Zoll observe that ‘initial onset replacement . . . is probably the most common type of case discussed in the literature’ (2005, p. 42). This sets it apart from ‘fixed-segmentism’ in child templates, which affects the onset of the second consonant in a word form, not the first. We will briefly elaborate upon this point.

To date no templates have been identified with the word-initial consonant as a fixed element alone in child templatic patterns (i.e., aside from consonantal melodies like that of Macken’s Si, which affect both onsets of a disyllabic word form). It is only C2 that tends to be fixed: see, for example, Table 3, where medial yod is the fixed element (see the matrix of reported template melodies, Vihman, 2014, table 10.2, pp. 300f.). This may relate to the fact that word-initial position is the easiest to perceive and remember, especially for a strongly trochaic language like English with no medial geminates (see the discussion of the special status of the word-initial consonant in Vihman, Nakai, DePaolis, & Hallé, 2004). However, the principle also holds for French (for example, see the medial [l]-based template of a French child; Vihman, 1993). The initial consonant is also the easiest to plan for in production, whereas use of a fixed consonant in medial position can reduce the memory load by providing an anchor for mapping adult targets to a production routine.

8 Relation of Child Phonological Templates to Compound Formation and Irreversible Binomials

A very different source of adult-language parallels may be looked for in the marginal area of grammar constituted by ‘reduplicative compounds’, a formation that has recently begun to gain more attention, sometimes in a cross-linguistic framework (e.g., for ‘echo-pairs’, see Sóskuthy, 2012; see also Keane, 2001; Southern, 2005, cited in Sóskuthy, 2012). With specific reference to English the term ‘rhyming compounds’ is used for the phenomenon we are discussing here (Katamba, 2005, cited in Benczes, 2012). Benczes (2013) has provided examples of innovative uses of novel compounds in English that draw on alliteration and rhyme in the selection of their component word forms (e.g., wall wart, cuddle bubble), arguing that the phonological features in question render novel coinages not only aesthetically pleasing and entertaining but also memorable—the latter in particular being a critical function of child templates as well.

Table 11 provides examples of English rhyming compounds in current use. Although some such compounds consist of two known words (e.g., fender-bender, snail mail), most examples are made up of a recognizable word and a rhyming copy (Holy Moley) or two meaningless units combined to create a meaningful compound (boogie-woogie).

Table 11. English rhyming compounds

Labial in 2d element of compound

Non-labial in 2d element

easy-peasy

fuddy-duddy

hanky-panky

heebie-jeebies

hodge-podge

helter-skelter

namby-pamby

inky-dinky

roly-poly

nitty-gritty

teen(s)y-ween(s)y

razzle-dazzle

Note: (*) italics indicate forms in which both elements are identifiable words still in current use with the meaning that appears to obtain in the compound).

Analysis of an informal collection of 97 such forms shows a strong bias toward a labial onset in the second element (48%, or 47/97); a bias toward use of /h/as onset of the first word can be observed as well (32, or 33% of the same collection). Benczes (2012) reports 39% h-initial words in Thun’s (1963) collection of about 2,000 compounds, and in a historical account reveals that h-initial reduplicative compounds were the earliest to be recorded (15th century), often with a meaning related to ‘disorder of some sort’; some of the very first examples were associated with magic (e.g., hocus-pocus, in contemporary usage). Benczes suggests that these forms may have constituted the model from which a schema (or template) was extracted, which we might state as <h[. . .]1C2[. . .]1>—i.e., two rhyming components, the first with onset/h/, the second with some other consonant—or alternatively, <C[ . .]1[lab C][. . .]1>, to accommodate the larger proportion of compounds with labial onset to the second element. These patterns resemble children’s templates in being purely lexical, involving no grammatical meaning, although they are formally more complex, besides reflecting the unconscious work of many speakers over centuries. Furthermore, the occurrence of labials as second-element onset in nearly half of the compounds conforms with the finding that in individual child templatic patterns it is only second-syllable onsets or word-final codas, not word-initial consonants, that tend to be fixed.

Interestingly, Sóskuthy’s (2012) systematic analysis of the phonological constraints on Hungarian echo-words obtained from a Web-based corpus reveals that, in those data too, labials predominate as onset to the second element. Similarly, Inkelas (2003) describes a reduplication game invented by her son J. that lasted from age 2;5 to 4;9; here again, labials initially formed the onset of the reduplicants supplied by the child, which copied the final stress-foot of a range of English words provided by the parent-investigators. Inkelas also provides examples from Turkish and Abkhaz sources of adult-language reduplicative patterns with fixed labial onset for the copy-element. This striking (though partial) regularity has not yet been accounted for.

In addition to rhyming compounds of the kind listed in Table 11, English partial reduplication encompasses large numbers of what have been called ‘ablaut-motivated’ compounds, or compounds whose two elements are phonologically related through alliteration rather than rhyme (Bauer, 1983, cited in Benczes, 2012): these words virtually all seem to follow the <[high]. . .[low]> vocalic sequence familiar in English from past tense alternations like sing: sang (cf. fiddle-faddle, flim-flam, but also criss-cross, sing-song, wishy-washy—and doodad, gewgaw). This pattern is the opposite of the one vocalic melody (<[low]. . .[high]>) sometimes found in child phonological templates.

Finally, phonological effects that shape semi-formulaic phrases can also be identified as playing a role like that of child templates. Malkiel (1959), in an in-depth account of ‘irreversible binomials’ (or frequent word sequences that rarely occur in the opposite order) that draws on several European languages, points out the prevalence among them of rhyme (by hook or by crook, town and gown, French (n’avoir) ni feu ni lieu ‘(to have) neither hearth nor home [lit. ‘neither fire nor place]’, Spanish de tomo y lomo ‘bulky and heavy, consequential’, Russian ni dat’ ni v’at’ ‘just so [lit. ‘neither give nor take away’],’ as well as assonance (hit or miss, rise and shine). Another such effect is the rhythmic principle, short before long, which may be further supported by alliteration (e.g., judge and jury, [without] rhyme or reason, rough and ready, whys and wherefores). Malkiel finds alliteration to be the dominant ‘soldering device’ in Germanic languages (1959, p. 123). Of the 200-odd English pairings of words he cites that differ in length, the second element is longer in 166 (82%). Of these <short-long> binomials, 58 (35%) also show alliteration. Thus the combination of fixed accentual pattern and alliteration (or ‘consonant harmony’) that is so characteristic of child templates also finds its parallel here, in the lexicalization of commonly co-occurring collocations.

9 Concluding Thoughts: Templates in Early Vocabulary Development

We have illustrated many different types of templates identifiable in children’s early word forms, noting variability from one child to the next but also some general principles or constraints. The chief common limitation is to a length of no more than one or two syllables, which seems to be the longest sequence that a child at this developmental stage can represent well enough to permit voluntary access. We also found examples of templatic structures that occur in some language types but not others; the key ambient-language factor shaping child phonological templates appears to be accentual structure, including both vowel and consonant duration as well as stress pattern.

Development of a somewhat flexible template is assumed to facilitate the acquisition of new words in a period of intense lexical learning. The development of one or more favored patterns permits rapid mapping and assimilation of novel sequences, which can then be retained well enough for later production, even without priming or contextual support. We provided a speculative account of the origins of templates in the combination of item learning, with attention, and subsequent pattern generalization, or secondary distributional learning.

Finally, we considered some possible parallels in adult language use. Although templates play a role in contemporary phonological theory, notably in non-concatenative morphology, we define ‘template’ somewhat differently here; reduplication with fixed segmentism also shares some but not all of the formal characteristics of child templates. The association with the expression of grammatical meaning in either of those domains is quite different from the purely lexical—and transient—nature of child phonological templates.

Instead, the closest parallel to the role played by templates in children as we have observed them may be in productive areas of lexical composition, such as rhythmic compounds, a source of novel word formation documented for English for at least 600 years, echo words in other languages (such as Hungarian and Kamrupi), and irreversible binomials, also a cross-linguistic phenomenon. Although the rhythmic compounds fall into a small number of different semantic categories—suggesting possible sound-symbolic value for adults in the rhythmic compound as such, quite unlike the function of templates in child use—for both compounds and binomials it is the inherent aesthetic appeal and memorability of repeating phonological elements (whether full rhyme, alliteration, or assonance) that seems to hold the key to the longevity of the forms so created and the continuing function of these processes in generating more neologisms and lexicalized expressions. Infants do not learn templates from the input, nor do adult templates derive from those of infants, and rhyme itself is not a notable element in child templates. Nevertheless, infants and adults appear to be drawing, for different purposes, on their common resources in human sensitivity to repetition as a support for memory.

Further Reading

Bybee, J. (2006). From usage to grammar: The mind’s response to repetition. Language, 82, 711–733.Find this resource:

Ellis, N. C. (2002). Frequency effects in language processing: A review with implications for theories of implicit and explicit language acquisition. Studies in Second Language Acquisition, 24, 143–188.Find this resource:

Lindsey, S., & Gaskell, M. G. (2010). A complementary systems account of word learning in L1 and L2. Language Learning, 60(Suppl. 2), 45–63.Find this resource:

Menn, L., Schmidt, E., & Nicholas, B. (2013). Challenges to theories, charges to a model: The linked-attractor model of phonological development. In M. M. Vihman & T. Keren-Portnoy (Eds.), The emergence of phonology: Whole word approaches, cross-linguistic evidence. Cambridge, U.K.: Cambridge University Press.Find this resource:

Munson, B., Edwards, J., & Beckman, M. E. (2012). Phonological representation in language acquisition: Climbing the ladder of abstraction. In A. C. Cohn, C. Fougeron, & M. K. Huffman (Eds.), Oxford handbook of laboratory phonology (pp. 288–309). Oxford: Oxford University Press.Find this resource:

Pierrehumbert, J. (2003a). Phonetic diversity, statistical learning, and acquisition of phonology. Language and Speech, 46, 115–154.Find this resource:

Swingley, D. (2009). Contributions of infant word learning to language development. Philosophical Transactions of the Royal Society, B, 364, 3617–3632.Find this resource:

Vihman, M. M. (2014). Phonological development: The first two years (2d ed.). Malden, MA: Wiley-Blackwell.Find this resource:

Vihman, M. M. (2016). Prosodic structures and templates in bilingual phonological development. Bilingualism: Language and Cognition, 19, 69–88.Find this resource:

Vihman, M. M., & Croft, W. (2007). Phonological development: Toward a ‘radical’ templatic phonology. Linguistics, 45, 683–725.Find this resource:

Vihman, M. M., DePaolis, R. A., & Keren-Portnoy, T. (2014). The role of production in infant word learning. Language Learning, 64(Suppl. 2), 121–140.Find this resource:

Vihman, M. M., DePaolis, R. A., & Keren-Portnoy, T. (2015). A dynamic systems approach to babbling and words. In E. Bavin (Ed.), Handbook of child language (2d ed., pp. 207–229). Cambridge, U.K.: Cambridge University Press.Find this resource:

Vihman, M. M., & Keren-Portnoy, T. (Eds.). (2013). The emergence of phonology: Whole word approaches, cross-linguistic evidence. Cambridge, U.K.: Cambridge University Press.Find this resource:

Vihman, M. M., & Wauquier, S. (in press). Templates in child language. In M. Hickmann, E. Veneziano, & H. Jisa (Eds.), Sources of variation in first language acquisition: Languages, contexts, and learners. Amsterdam: John Benjamins.Find this resource:

References

Alderete, J., Beckman, J., Benua, L., Gnanadesikan, A., McCarthy, J. J., & Urbanczyk, S. (1999). Reduplication with fixed segmentism. Linguistic Inquiry, 30, 327–364.Find this resource:

Anwaar, H. (2015). Phonological development of a bilingual child (Unpublished). University of York.Find this resource:

Baia, M. F. A. (2013). Os templates no desenvolvimento fonológico: O caso do Português Brasileiro (Unpublished PhD diss.). University of São Paulo, Brazil.Find this resource:

Barlow, M., & Kemmer, S. (Eds.). (2000). Usage-based models of language. Stanford, CA: CSLI Publications.Find this resource:

Bat-El, O. (2010). The template is not a unit: The minimal word in adult and child Hebrew. Talk presented at the 7th Old World Conference in Phonology, Nice, January.Find this resource:

Bauer, I. (1983). English word-formation. Cambridge, U.K.: Cambridge University Press.Find this resource:

Bauer, P. J. (2008). Toward a neuro-developmental account of the development of declarative memory. Psychobiology, 50, 19–31.Find this resource:

Becker, M., & Tessier, A. M. (2011). Trajectories of faithfulness in child-specific phonology. Phonology, 28, 163–196.Find this resource:

Benczes, R. (2012). Just a load of hibber-gibber? Making sense of English rhyming compounds. Australian Journal of Linguistics, 32, 299–326.Find this resource:

Benczes, R. (2013). The role of alliteration and rhyme in novel metaphorical and metonymical compounds. Metaphor and Symbol, 28, 167–184.Find this resource:

Berent, I., Marcus, G. F., Shimron, J., & Gafos, A. I. (2002). The scope of linguistic generalizations: Evidence from Hebrew word formation. Cognition, 83, 113–139.Find this resource:

Berg, T. (1992). Phonological harmony as a processing problem. Journal of Child Language, 19, 225–257.Find this resource:

Berg, T., & Schade, U. (2000). A local connectionist account of consonant harmony in child language. Cognitive Science, 24, 123–149.Find this resource:

Bergelson, E., & Swingley, D. (2012). At 6 to 9 months, human infants know the meanings of many common nouns. Proceedings of the National Academy of Sciences of the USA, 109, 3253–3258.Find this resource:

Bergelson, E., & Swingley, D. (2013). The acquisition of abstract words. Cognition, 127, 391–397lFind this resource:

Bernhardt, B., & Stemberger, J. (1998). Handbook of phonological development from the perspective of constraint-based nonlinear phonology. San Diego, CA: Academic Press.Find this resource:

Bhaya Nair, R. (1991). Monosyllabic English or disyllabic Hindi? Indian Linguistics, 52, 51–90.Find this resource:

Bortfeld, H., Morgan, J. L., Golinkoff, R. M., & Rathbun, K. (2005). Mommy and me: Familiar names help launch babies into speech stream segmentation. Psychological Science, 16, 298–304.Find this resource:

Brent, M. R., & Siskind, J. M. (2001). The role of exposure to isolated words in early vocabulary development. Cognition, 81, B33–B44.Find this resource:

Brulard, I., & Carr, P. (2003). Phonological templates and strategies in French/English bilingual acquisition. International Journal of Bilingualism, 7, 177–202.Find this resource:

Bybee, J. (2001). Phonology and language use. Cambridge, U.K.: Cambridge University Press.Find this resource:

Bybee, J. (2006). From usage to grammar: The mind’s response to repetition. Language, 82, 711–733.Find this resource:

Bybee, J., & Hopper, P. J. (Eds.). (2001). Frequency and the emergence of linguistic structure. Amsterdam: John Benjamins.Find this resource:

Bybee, J. L., & Slobin, D. I. (1982). Rules and schemas in the development and use of the English past tense. Language, 58, 265–289.Find this resource:

Carey, S. (1978). The child as word learner. In M. Halle, J. Bresnan, & G. A. Miller (Eds.), Linguistic Theory and Psychological Reality (pp. 264–293). Cambridge, MA: MIT Press.Find this resource:

Carvalho, J. B. de (2004). Templatic morphology in the Portuguese verb. In T. Meisenburg & M. Selig (Eds.), Nouveaux départs en phonologie: les conceptions sub- et suprasegmentales (pp. 13–32). Tübingen, Germany: Gunter Narr Verlag.Find this resource:

Chomsky, N., & Halle, M. (1968). The sound pattern of English. New York: Harper & Row.Find this resource:

Dawdy-Hesterberg, L. G., & Pierrehumbert, J. B. (2014). Learnability and generalisation of Arabic broken plural nouns. Language, Cognition and Neuroscience, 29, 1268–1282.Find this resource:

DePaolis, R., Vihman, M. M., & Keren-Portnoy, T. (2011). Do production patterns influence the processing of speech in prelinguistic infants? Infant Behavior and Development, 34, 590–601.Find this resource:

DePaolis, R., Vihman, M. M., & Nakai, S. (2013). The influence of babbling patterns on the processing of speech. Infant Behavior and Development, 36, 642–649.Find this resource:

Elbers, L., & Wijnen, F. (1992). Effort, production skill, and language learning. In C. A. Ferguson, L. Menn, & C. Stoel-Gammon (Eds.), Phonological development: Models, research, implications. Timonium, MD: York Press.Find this resource:

Ellis, N. C. (2002). Frequency effects in language processing: A review with implications for theories of implicit and explicit language acquisition. Studies in Second Language Acquisition, 24, 143–188.Find this resource:

Ellis, N. C. (2005). At the interface: Dynamic interactions of explicit and implicit language knowledge. Studies in Second Language Acquisition, 27, 305–352.Find this resource:

Elsen, H. (1996). Two routes to language. First Language, 16, 141–158.Find this resource:

Ferguson, C. A., & Farwell, C. B. (1975). Words and sounds in early language acquisition. Language, 51, 419–439.Find this resource:

Fikkert, P., & Levelt, C. (2008). How does Place fall into place? The lexicon and emergent constraints in children’s developing grammars. In P. Avery, E. Dresher, & K. Rice (Eds), Contrast in phonology: Theory, perception, acquisition (pp. 231–270). Berlin: Mouton.Find this resource:

Ganger, J., & Brent, M. (2004). Reexamining the vocabulary spurt. Developmental Psychology, 40, 621–632.Find this resource:

Gathercole, S. E., & Baddeley, A. D. (1993). Working memory and language. Mahwah, NJ: Erlbaum.Find this resource:

Gershkoff-Stowe, L. (2002). Object naming, vocabulary growth, and the development of word retrieval abilities Journal of Memory and Language, 46, 665–687.Find this resource:

Gnanadesikan, A. (2004). Markedness and faithfulness constraints in child phonology. In R. Kager, J. Pater, & W. Zonneveld (Eds.), Constraints in phonological acquisition (pp. 73–108). Cambridge, U.K.: Cambridge University Press.Find this resource:

Goad, H. (1997). Consonant harmony in child language: An optimality theoretic account. In S. J. Hannahs & M. Young-Scholten (Eds.), Focus on phonological acquisition (pp. 113–142). Amsterdam: John Benjamins.Find this resource:

Goswami, U. (1955–1956). Onomatopoetic words and echo-words in Kamrupi. Indian Linguistics, 17, 161–164.Find this resource:

Hansson, G. Ó. (2010). Consonant harmony: Long-distance interaction in phonology. Berkeley, CA: University of California Press.Find this resource:

Huber, R., & Born, J. (2013). Sleep, synaptic connectivity and hippocampal memory during early development. Trends in Cognitive Sciences, 18, 141–152.Find this resource:

Icht, M., & Mama, Y. (2015). The production effect in memory: A prominent mnemonic in children. Journal of Child Language, 42, 1102–1124.Find this resource:

Inkelas, S. (2003). J’s rhymes: A longitudinal case study of language play. Journal of Child Language, 30, 557–581.Find this resource:

Inkelas, S., & Zoll, C. (2005). Reduplication: Doubling in morphology. Cambridge, U.K.: Cambridge University Press.Find this resource:

Johnson, E. K. (2012). Bootstrapping language: Are infant statisticians up to the job? In P. Rebuschat & J. Williams (Eds.), Statistical learning and language acquisition (pp. 55–90). Berlin: Mouton de Gruyter.Find this resource:

Jusczyk, P. W. (1997). The discovery of spoken language. Cambridge, MA: MIT Press.Find this resource:

Katamba, F. (2005). English words: Structure, history, usage (2d ed.). London: Routledge.Find this resource:

Keane, E. (2001). Echo words in Tamil (Unpublished PhD thesis). Oxford University.Find this resource:

Kehoe, M. M. (2015). Lexical-phonological interactions in bilingual children. First Language, 35, 93–125.Find this resource:

Kelly, M. H., & Martin, S. (1994). Domain-general abilities applied to domain-specific tasks: Sensitivity to probabilities in perception, cognition, and language. Lingua, 92, 105–140.Find this resource:

Keren-Portnoy, T., Majorano, M., & Vihman, M. M. (2009). From phonetics to phonology: The emergence of first words in Italian. Journal of Child Language, 36, 235–267.Find this resource:

Keren-Portnoy, T., & Segal, O. (in press). Phonological development in Hebrew- learning infants and toddlers: perception and production. In R. Berman (Ed.), Acquisition of Hebrew from infancy to adolescence. Amsterdam: John Benjamins.Find this resource:

Keren-Portnoy, T., Vihman, M. M., DePaolis, R., Whitaker, C., & Williams, N. A. (2010). The role of vocal practice in constructing phonological working memory. Journal of Speech, Language, and Hearing Research, 53, 1280–1293.Find this resource:

Khattab, G., & Al-Tamimi, J. (2013). Early phonological patterns in Lebanese Arabic. In M. M. Vihman & T. Keren-Portnoy (Eds.), The emergence of phonology: Whole word approaches, cross-linguistic evidence (pp. 374–414). Cambridge, U.K.: Cambridge University Press.Find this resource:

Kirkham, N. Z., Slemmer, J. A., & Johnson, S. P. (2002). Visual statistical learning in infancy: Evidence for a domain-general learning mechanism. Cognition, 83, B35–B42.Find this resource:

Kuhl, P. K., Conboy, B. T., Coffey-Corina, S., Padden, D., Rivera-Graxiola, M., & Nelson, T. (2008). Phonetic learning as a pathway to language: New data and native language magnet theory expanded (NLM-e). Philosophical Transactions of the Royal Society B-Biological Sciences, 363(1493), 979–1000.Find this resource:

Kunnari, S. (2000). Characteristics of early lexical and phonological development in children acquiring Finnish. Acta Universitatis Ouluensis B34, University of Oulu.Find this resource:

Laing, C. (2016). ‘What does the cow say?’ An exploratory analysis of onomatopoeia in early phonological development (Unpublished PhD thesis). University of York.Find this resource:

Lew-Williams, C., Pelucchi, B., & Saffran, J. R. (2011). Isolated words enhance statistical learning by 9-month-old infants. Developmental Science, 1–7.Find this resource:

Lin, Y., & Mielke, J. (2008). Discovering place and manner features—What can be learned from acoustic and articulatory data? In J. Tauberer, A. Eilam, & L. MacKenzie (Eds.), Penn Working Papers in Linguistics, 14, 241–254.Find this resource:

Lindsey, S., & Gaskell, M. G. (2010). A complementary systems account of word learning in L1 and L2. Language Learning, 60(Suppl. 2), 45–63.Find this resource:

Lleó, C. (1990). Homonymy and reduplication: On the extended availability of two strategies in phonological acquisition. Journal of Child Language, 17, 267–278.Find this resource:

Macken, M. A. (1978). Permitted complexity in phonological development: One child’s acquisition of Spanish consonants. Lingua, 44, 219–253.Find this resource:

Macken, M. A. (1979). Developmental reorganization of phonology: A hierarchy of basic units of acquisition. Lingua, 49, 11–49.Find this resource:

Macken, M. A. (1995). Phonological acquisition. In J. A. Goldsmith (Ed.), The Handbook of Phonological Theory (pp. 671–696). Oxford: Blackwell.Find this resource:

Mackenzie, S. (2009). Contrast and similarity in consonant harmony processes (Unpublished PhD thesis). University of Toronto.Find this resource:

MacLeod, C., Gopie, N., Hourihan, K. L., Neary, K. R., & Ozubko, K. D. (2010). The production effect: Delineation of a phenomenon. Journal of Experimental Psychology: Learning, Memory, and Cognition, 36, 671–685.Find this resource:

Majorano, M., Vihman, M. M., & DePaolis, R. A. (2014). The relationship between infants’ production experience and their processing of speech. Language Learning and Development, 10, 179–204.Find this resource:

Malkiel, Y. (1959). Studies in irreversible binomials. Lingua, 8, 113–160.Find this resource:

Maye, J., Werker, J. F., & Gerken, L. (2002). Infant sensitivity to distributional information can affect phonetic discrimination. Cognition, 82, B101–B111.Find this resource:

McAllister Byun, T., & Inkelas, S. (2014). Child consonant harmony and phonologization of performance errors. In H-L. Huang, E. Poole, & A. Rysling (Eds.), Proceedings of NELS 43 (pp. 291–302). Amherst, MA: GLSA.Find this resource:

McCarthy, J. J. (1981). A prosodic theory of nonconcatenative morphology. Linguistic Inquiry, 12, 373–418.Find this resource:

McCarthy, J. (1982). Prosodic templates, morphemic templates, and morphemic tiers. In H. van der Hulst & N. Smith (Eds.), The structure of phonological representations, I (pp. 191–223). Dordrecht, The Netherlands: Foris Publications.Find this resource:

McCarthy, J. J., & Prince, A. S. (1995). Prosodic morphology 1. In J. A. Goldsmith (Ed.), The Handbook of Phonological Theory (pp. 318–366). Oxford: Blackwell.Find this resource:

McClelland, J. L., McNaughton, B. L., & O’Reilly, R. C. (1995). Why there are complementary learning systems in the hippocampus and neocortex: Insights from the successes and failures of connectionist models of learning and memory. Psychological Review, 102, 419–517.Find this resource:

McCune, L., & Vihman, M. M. (2001). Early phonetic and lexical development. Journal of Speech, Language and Hearing Research, 44, 670–684.Find this resource:

McMurray, B., & Hollich, G. (2009). Core computational principles of language acquisition: Can statistical learning do the job? Introduction to special section. Developmental Science, 12, 365–368.Find this resource:

Menn, L. (1971). Phonotactic rules in beginning speech: A study in the development of English discourse. Lingua, 26, 225–251.Find this resource:

Menn, L. (1983). Development of articulatory, phonetic, and phonological capabilities. In B. Butterworth (Ed.), Language Production, Vol. 2 (pp. 3–50). London: Academic Press. Reprinted in Vihman & Keren-Portnoy (2013).Find this resource:

Menn, L., Schmidt, E., & Nicholas, B. (2013). Challenges to theories, charges to a model: The linked-attractor model of phonological development. In M. M. Vihman & T. Keren-Portnoy (Eds.), The emergence of phonology: Whole word approaches, cross-linguistic evidence (pp. 460–502). Cambridge, U.K.: Cambridge University Press.Find this resource:

Menn, L., & Vihman, M. M. (2011). Features in child phonology: Inherent, emergent, or artefacts of analysis? In N. Clements & R. Ridouane (Eds.), Where do phonological features come from? The nature and sources of phonological primitives (pp. 261–301). Amsterdam: John Benjamins.Find this resource:

Merhav, M., Karni, A., & Gilboa, A. (2015). Not all declarative memories are created equal: Fast mapping as a direct route to cortical declarative representations. NeuroImage, 117, 80–92.Find this resource:

Munson, B., Edwards, J., & Beckman, M. E. (2012). Phonological representation in language acquisition: Climbing the ladder of abstraction. In A. C. Cohn, C. Fougeron, & M. K. Huffman (Eds.), Oxford Handbook of Laboratory Phonology (pp. 288–309). Oxford: Oxford University Press.Find this resource:

Oliveira-Guimarães, D. (2013). Beyond early words: Word template development in Brazilian Portuguese. In M. M. Vihman & T. Keren-Portnoy (Eds.), The emergence of phonology: Whole-word approaches, cross-linguistic evidence (pp. 291–316). Cambridge, U.K.: Cambridge University Press.Find this resource:

Parladé, M. V., & Iverson, J. M. (2011). The interplay between language, gesture, and affect during communicative transition: A dynamic systems approach. Developmental Psychology, 47, 820–833.Find this resource:

Pater, J., & Werle, A. (2003). Direction of assimilation in child consonant harmony. Canadian Journal of Linguistics, 48, 385–408.Find this resource:

Pierrehumbert, J. (2003). Phonetic diversity, statistical learning, and acquisition of phonology. Language and Speech, 46, 115–154.Find this resource:

Priestly, T. M. S. (1977). One idiosyncratic strategy in the acquisition of phonology. Journal of Child Language, 4, 45–66.Find this resource:

Renner, L., & Strandberg, A. (2015). Phonological templates in Swedish 18-month-old children in relation to vocabulary size. Poster presented at Child Language Symposium, Warwick, July.Find this resource:

Saffran, J. R., Aslin, R. N., & Newport, E. L. (1996). Statistical learning by 8-month-old infants. Science, 274, 1926–1928.Find this resource:

Saffran, J. R., Newport, E. L., Aslin, R. N., Tunick, R. A., & Barrueco, S. (1997). Incidental language learning: Listening (and learning) out of the corner of your ear. Psychological Science, 8, 101–105.Find this resource:

Savinainen-Makkonen, T. (2007). Geminate template: A model for first Finnish words. First Language, 27, 347–359.Find this resource:

Scheer, T. (2003). The key to Czech vowel length: Templates. In P. Kosta, J. Blaszczak, J. Frasek, L. Geist, & M. Zygis (Eds.), Investigations into Formal Slavic Linguistics (pp. 97–118). Frankfurt: Lang.Find this resource:

Sharon, T., Moscovitch, M., & Gilboa, A. (2011). Rapid neocortical acquisition of long-term arbitrary associations independent of the hippocampus. Proceedings of the National Academy of Sciences, 117, 80–92.Find this resource:

Shaw, P. A. (1991). Consonant harmony systems. In C. Paradis & J-F. Prunet (Eds.), Phonetics and phonology: The special status of coronals, Vol. 2 (pp. 125–157). San Diego, CA: Academic Press.Find this resource:

Shimron, J. (Ed.). (2003). Language processing and acquisition in languages of Semitic, root-based morphology. Amsterdam: John Benjamins.Find this resource:

Smith, N. V. (1973). The acquisition of phonology: A case study. Cambridge, U.K.: Cambridge University Press.Find this resource:

Sóskuthy, M. (2012). Morphology in the extreme: Echo-pairs in Hungarian. In F. Kiefer & Z. Bánréti (Eds.), Twenty years of theoretical linguistics in Budapest (pp. 123–144). Budapest: Tinta Publishing House.Find this resource:

Southern, M. R. V. (2005). Contagious couplings: Transmission of expressives in Yiddish echo phrases. London: Praeger.Find this resource:

Stager, C. L., & Werker, J. F. (1997). Infants listen for more phonetic detail in speech perception than in word-learning tasks. Nature, 388, 381–382.Find this resource:

Stoel-Gammon, C., & Cooper, J. A. (1984). Patterns of early lexical and phonological development. Journal of Child Language, 11, 247–271.Find this resource:

Stoel-Gammon, C., & Stemberger, J. P. (1994). Consonant harmony and phonological underspecification in child speech. In M. Yavas (Ed.), First and second language phonology (pp. 63–80). San Diego, CA: Singular.Find this resource:

Swingley, D. (2009). Contributions of infant word learning to language development. Philosophical Transactions of the Royal Society, B, 364, 3617–3632.Find this resource:

Szreder, M. (2013). The acquisition of consonant clusters in Polish: A case study. In M. M. Vihman & T. Keren-Portnoy (Eds.), The emergence of phonology: Whole-word approaches, cross-linguistic evidence. Cambridge, U.K.: Cambridge University Press.Find this resource:

Tesar, B. (2013). Output-driven phonology. Cambridge, U.K.: Cambridge University Press.Find this resource:

Tessier, A-M. (2016). Phonological acquisition: Child language and constraint-based grammar. New York: Palgrave.Find this resource:

Thun, N. (1963). Reduplicative words in English: A study of formations of the types tick-tick, hurly-burly and shilly-shally (Unpublished PhD thesis). Uppsala University, Uppsala.Find this resource:

Velten, H. V. (1943). The growth of phonemic and lexical patterns in infant language. Language, 19, 231–292.Find this resource:

Veneziano, E., & Sinclair, H. (2000). The changing status of ‘filler syllables’ on the way to grammatical morphemes. Journal of Child Language, 27, 461–500.Find this resource:

Vihman, M. M. (1976). From prespeech to speech: On early phonology. Stanford Papers and Reports on Child Language Development, 12, 230–244.Find this resource:

Vihman, M. M. (1978). Consonant harmony: Its scope and function in child language. In J. Greenberg, C. A. Ferguson, & E. A. Moravcsik (Eds.), Universals of Human Language, 2 (pp. 281–334). Stanford, CA: Stanford University Press.Find this resource:

Vihman, M. M. (1981). Phonology and the development of the lexicon: Evidence from children’s errors. Journal of Child Language, 8, 239–264.Find this resource:

Vihman, M. M. (1993). Variable paths to early word production. Journal of Phonetics, 21, 61–82.Find this resource:

Vihman, M. M. (2014). Phonological development: The first two years (2d ed.). Malden, MA: Wiley-Blackwell.Find this resource:

Vihman, M. M. (2015). Perception and production in phonological development. In B. MacWhinney & W. O’Grady (Eds.), Handbook of language emergence (pp. 437–457). Malden, MA: Wiley-Blackwell.Find this resource:

Vihman, M. M. (2016). Prosodic structures and templates in bilingual phonological development. Bilingualism: Language and Cognition, 19, 69–88.Find this resource:

Vihman, M. M. (in press). The development of prosodic structure: A usage-based approach. In P. Prieto & N. Esteve-Gilbert (Eds.), Prosodic development in first language acquisition. Amsterdam: John Benjamins.Find this resource:

Vihman, M. M., & Croft, W. (2007). Phonological development: Toward a ‘radical’ templatic phonology. Linguistics, 45, 683–725.Find this resource:

Vihman, M. M., DePaolis, R. A., & Keren-Portnoy, T. (2014). The role of production in infant word learning. Language Learning, 64(Suppl. 2), 121–140.Find this resource:

Vihman, M. M., & Keren-Portnoy, T. (Eds.). (2013). The emergence of phonology: Whole word approaches, cross-linguistic evidence. Cambridge, U.K.: Cambridge University Press.Find this resource:

Vihman, M. M., Keren-Portnoy, T., Whitaker, C., Bidgood, A., & McGillion, M. (2013). Late-talking toddlers: Relating early phonological development to later language advance. York Papers in Linguistics, 13, 47–69.Find this resource:

Vihman, M. M., & Kunnari, S. (2006). The sources of phonological knowledge: A cross-linguistic perspective. Recherches Linguistiques de Vincennes, 35, 133–164.Find this resource:

Vihman, M. M., & Majorano, M. (in press). The role of geminates in infants’ early words and word-form recognition. Journal of Child Language.Find this resource:

Vihman, M. M., Nakai, S., DePaolis, R. A., & Hallé, P. (2004). The role of accentual pattern in early lexical representation. Journal of Memory and Language, 50, 336–353.Find this resource:

Vihman, M. M., & Velleman, S. (1989). Phonological reorganization: A case study. Language and Speech, 32, 149–170.Find this resource:

Vihman, M. M., & Velleman, S. L. (2000). Phonetics and the origins of phonology. In N. Burton-Roberts, P. Carr, & G. Docherty (Eds.), Phonological knowledge: Its nature and status (pp. 305–339). Oxford: Oxford University Press.Find this resource:

Vihman, M., Velleman, S., & McCune, L. (1994). How abstract is child phonology? In M. Yavas (Ed.), First and second language phonology (pp. 9–44). San Diego, CA: Singular.Find this resource:

Vihman, M. M., & Vihman, V-A. (2011). From first words to segments: A case study in phonological development. In I. Arnon & E. V. Clark (Eds.), Experience, variation, and generalization: Learning a first language (pp. 109–133). Amsterdam: John Benjamins.Find this resource:

Vihman, M. M., & Wauquier, S. (in press). Templates in child language. In M. Hickmann, E. Veneziano, & H. Jisa (Eds.), Sources of variation in first language acquisition: Languages, contexts, and learners. Amsterdam: John Benjamins.Find this resource:

Waterson, N. (1971). Child phonology: A prosodic view. Journal of Linguistics, 7, 179–211.Find this resource:

Wauquier, S. & Yamaguchi, N. (2013). Templates in French. In M. M. Vihman & T. Keren-Portnoy (Eds.), The emergence of phonology: Whole-word approaches, cross-linguistic evidence (pp. 317–342). Cambridge, U.K.: Cambridge University Press.Find this resource:

Werker, J. F., & Curtin, S. (2005). PRIMIR: A developmental framework of infant speech processing. Language, Learning and Development, 1, 197–234.Find this resource:

Zamuner, T. S., Morin-Lessard, E., Strahm, S., & Page, M. P. A. (2016). Spoken word recognition of novel words, either produced or only heard during learning. Journal of Memory and Language, 89, 55–67.Find this resource:

Notes:

(1.) Arabic constitutes a notable exception (Khattab & Al Tamimi, 2013): Although geminates are common word-medially and are picked up and used early by the children, ‘null’ (vowel) onsets are rare, presumably due to shaping by the adult language, whose structure requires provision of a glottal stop wherever full consonantal onset is absent.