Werker provides a compelling story for how infants’ speech perceptual abilities develop and change over the first years of life. Less well understood is the relationship between perception and production abilities. Given pervasive evidence of individual differences in production, this raises many questions about possible individual differences in perception as well, and how this might influence the course of language acquisition more generally. This Commentary reviews some of the types of individual differences found in production, some of what we know about how these are linked to perception, and areas for further research.
It has long been known that children follow certain general patterns when acquiring their first words. However, there are also individual differences, in the timing of when children begin to produce their first words as well as in the actual phonological shape that these early words take. Thus, many children use their first words by 18 months, with a word spurt around 50 words, reaching 300 words by 24 months (cf., Fenson, Dale, Reznick, Bates, & Thal, Reference Fenson, Dale, Reznick, Bates and Thal1994). However, others start to use their first words as early as 10 or 11 months, and are using short sentences by 18 months (e.g., two of the children in the Providence Corpus; Demuth, Culbertson, & Alter, Reference Demuth, Culbertson and Alter2006). Others (“late talkers”) may only begin to use their first words after the age of 2, catching up around the age of 3 (cf., Rescorla, Roberts, & Dahlsgaard, Reference Rescorla, Roberts and Dahlsgaard1997). All the above is considered to be within the normal range for children with no obvious evidence of language disorders. Yet little is known about the possible differences in the timing or characteristics of children’s perceptual abilities, and if or how this might forecast these differences in production. Thus, although there are reports of perceptual sensitive at 6 months correlating with global measures of vocabulary size 1 or 2 years later (e.g., Tsao, Liu, & Kuhl, Reference Tsao, Liu and Kuhl2004), very little is known about the relationship between the perception and production of the same contrasts. For example, despite the fact that many children’s first words take the form of simple, “unmarked” segmental, syllable, and prosodic word structures (e.g., Demuth, Reference Demuth1995; Gnanadesikan, Reference Gnanadesikan2004), we know little about why this evolves differently depending on the child, even for a given language.
Thus, the field of child phonology, including children’s early word productions, appears to differ from that of the field of infant speech perception. However, this may be an artifact of the methods used. Studies of children’s phonological development have often taken the form of case studies, examining the speech productions of a few children, often longitudinally, providing an in-depth look at individual developmental patterns over time (e.g., Brown, Reference Brown1973). In contrast, infant speech perception studies have tended to involve groups of infants, reporting the development of language-specific perception abilities at a certain age. Although some theories of perception explicitly appeal to articulatory gestures in making claims about perception (e.g., Best’s, Reference Best1995, perceptual assimilation model), there has been little explicit research exploring the relationship between perception and production abilities in the same children. This is partly due to the different ages typically investigated, with much of the field of infant speech perception focusing on infants below the age of 2, and much of the child phonology literature focusing on children from the age of 2 onward. This, plus the different nature of the online versus speech production methods typically used, which necessitate the study of groups versus individuals, respectively, has led to little cross-talk between the two fields. Fortunately, this situation is slowly beginning to change, with concomitant advances in our theoretical understanding of the important connections between early perception and production.
In the following, I review some of what is known about individual differences in production, at the segmental level, interactions between segments and syllables, and at the level of prosodic words and prosodic morphology. I then briefly discuss recent research from our lab that begins to explore individual differences at the perception/production interface, discuss the possible mechanisms behind these individual differences, and identify areas for further research.
INDIVIDUAL DIFFERENCES IN THE PRODUCTION OF SEGMENTS, SYLLABLES, AND PROSODIC WORDS
Jakobson (Reference Jakobson1941) was one of the first to suggest that the segments children first use would be those that maximized contrasts. Thus, he proposed that the low vowel /a/ might be the first vowel to appear, along with the obstruent sonorant contrast /p/ versus /m/. Although the validity of Jakobson’s proposal has never been confirmed, many of these or similar segments do seem to be some of the first to appear in children’s early babbling and words, with consonants that are harder to produce, such as fricatives, affricates, and liquids later acquired (cf. Smit, Reference Smit1993). Thus, children’s early productions tend to include the more “unmarked” segments that occur widely across many languages. Of course, different languages have different phonological systems, and this also influences when and how phonemic contrasts are learned. For example, recent research on the early acquisition of Fulfulde (a West-Atlantic Niger–Congo language spoken in Mali and widely throughout West Africa) shows that, before 11 months, at least some children are using complex implosives and prenasal stops, and that implosives (e.g., /ɓ/) account for more than 37% of the obstruents one child produced (cf. Cissé, Demolin, & Vallée, Reference Cissé, Demolin and Vallée2011). It is not yet known if this was a particularly precocious child, or if this is a general pattern found for many Fulfulde-speaking children. It would then be interesting to know how and when perceptual sensitivity to, for example, implosive–nonimplosive contrasts emerges for such children.
Other more “marked” segments can also appear early in children’s speech. Dental clicks are produced early in languages like Xhosa (Gxilishe, Reference Gxilishe2004), whereas palatal alveolar clicks only appear later, and in Sesotho-speaking children’s words around the age of 3 (Demuth, Reference Demuth2003). One might expect perceptual sensitivity to all such consonants to appear long before the articulatory demands of production are met. The same might be thought to be true for segments like fricatives and liquids, where the articulatory challenges of producing such forms delays their appearance in children’s speech (cf. Lin & Demuth, Reference Lin and Demuth2015, for ultrasound study of English “l” development from 3–7 years).
Individual differences have long been documented at the segmental level in production, with some children preferring certain segments over others (Schwartz & Leaonard, Reference Schwartz and Leonard1982). For example, some have been shown to produce (many) words with labial–coronal sequences, or labial first words (cf. Yavaʂ, Reference Yavaʂ1995), depending on the language. Word truncation processes can also vary from child to child, where different segments may be preserved. When truncating trisyllabic words like banana to a disyllable, many children omit the first syllable altogether, producing a form like nana. Others, however, use the form bana, preserving the less sonorant (less marked) consonant in the onset position (cf. Demuth, Reference Demuth1996a; Pater, Reference Pater1997). Markedness in the selection of coda consonants also shows individual differences, with some English-speaking children’s first coda consonants being composed of less marked (more sonorant) codas (such as fricatives), whereas many others first use the most frequent (but more marked) coda consonant /t/ (Stites, Demuth, & Kirk, Reference Stites, Demuth and Kirk2004).
There are also many individual differences in the use of syllable structures and how these are produced. Some children begin using coda consonants with their first words, but many do not, producing predominantly reduplicated or consonant–vowel–consonant–vowel (CVCV) word forms for their first words (Demuth, Reference Demuth1995). A few, apparently precocious, children then go through a brief early stage of development (around 16 months) where they produce heavily released final (coda) consonants that can result in an “epenthetic” vowel, especially following voiced target codas (cf. Demuth et al., Reference Demuth, Culbertson and Alter2006; Vihman & Velleman Reference Vihman and Velleman2000). With respect to complex onsets, some children truncate word initial s-clusters by consistently dropping the /s/ (e.g., stop~top, slip~lip), whereas others preserve the more sonorant consonant, resulting in sop for stop, but lip for slip (cf. papers in Goad & Rose, Reference Goad and Rose2003). Individual differences in the use of consonant clusters are also found in Dutch-speaking children’s early words, with some showing earlier use of clusters in the onset (CCVC), but others first showing complexity in the coda (CVCC; Levelt, Schiller, & Levelt, Reference Levelt, Schiller and Levelt2000). Thus, there are many well-attested individual differences in both the timing and the realization of segments between children learning the same language. It is not clear if or how this might relate to perceptual differences as well. Little is known about individual differences in the perceptual sensitivity to such segmental/syllabic contrasts, and how this might relate to these individual differences in production.
Individual differences are also found in the form of children’s early prosodic words. For example, some Dutch-speaking children shift the placement of stress on words with a weak–strong (unstressed–stressed, iambic) stress pattern to a strong–weak (stressed–unstressed, trochaic) stress pattern, in keeping with the dominant stress pattern of the language (cf. Demuth, Reference Demuth1996b; Fikkert, Reference Fikkert1994). Although there has been much research on infant sensitivity to language-specific stress patterns (e.g., Jusczyk, Cutler, & Redanz, Reference Jusczyk, Cutler and Redanz1993; Yeung & Nazzi, Reference Yeung and Nazzi2014), little is known about individual differences in perception, and how this may relate to these differences in production.
Some children also tend to use “proto function words” often known as “fillers” (e.g., a wan a more “I want some more”), whereas others do not (cf. Peters, Reference Peters1983). It is unclear why this is the case, and if children’s perceptual sensitivities to function words (e.g., Shi, Werker, & Cutler, Reference Shi, Werker and Cutler2006) show concomitant individual differences. Some Spanish-speaking children actively truncate lexical items even while producing early “proto-articles” (amueka “la muñeca” (the doll)), while others never truncate lexical forms, only producing articles with fully formed words (cf. Demuth, Patrolia, Song, & Masapollo, Reference Demuth, Patrolia, Song and Masapollo2012). Individual differences are also found in English-speaking children’s use of articles, with most children prosodifying articles with the proceeding monosyllabic word (e.g., He [likes the] dolly; Gerken, Reference Gerken1996), but some actually stressing the article, initially producing it as its own prosodic word around the age of 20 months (He [likes] [the] dolly; cf. Demuth & McCullough, Reference Demuth and McCullough2009). Again, it is likely that perception of such forms is robust (cf. Gerken & McIntosh, Reference Gerken and McIntosh1993), but some of our resent research below suggests a close connection between perception and production.
INDIVIDUAL DIFFERENCES AT THE PERCEPTION/PRODUCTION INTERFACE
It is unclear if any of the individual differences in production discussed above correlate with differences in perception. However, there are indications that some of them might. For example, we have found that, in spontaneous speech, 2-year-olds tend to produce more inflectional morphemes, such as 3rd person singular –s, when these occur utterance finally compared to utterance medially (Song, Sundara, & Demuth, Reference Song, Sundara and Demuth2009). The final syllables of a phrase undergo phrase-final lengthening (Klatt, Reference Klatt1976; Lehiste, Reference Lehiste1973); we thus thought that phrase-final contexts might provided a longer durational context for all segments (and morphemes) to be produced. However, this phrase-final lengthening might also enhance the perceptual salience of these morphemes, making them easier to perceive as well. This was tested and found to be the case: both 20- and 27-month-olds showed sensitivity to 3rd person singular inflectional morphemes utterance finally, but not utterance medially, suggesting a close connection between perception and production (Sundara, Demuth, & Kuhl, Reference Sundara, Demuth and Kuhl2011). This finding also suggests that higher perceptual salience may facilitate “learning,” and this will then also be reflected earlier in production.
However, most verbs in English occur utterance medially, raising questions about when and how children begin to perceive and produce inflected verbs in this position. Further investigation revealed that those infants who showed higher 3rd person singular production abilities in an elicited imitation task also showed greater sensitivity to the presence versus (ungrammatical) absence of the morpheme utterance medially (e.g., She cries/cry now) during a listening task (Sundara et al., Reference Sundara, Demuth and Kuhl2011). Such findings suggest a tight correlation between perception and production at the level of the individual.
What could be the mechanism underlying these individual differences? Werker suggests that perceptual sensitivity interacts with meaning and referential context. We know that individual children’s receptive and expressive vocabulary sizes show large individual differences from the time these can be measured. Perhaps this is a result of more efficient perceptual development, leading to earlier production as well? Recent findings suggest that this may be the case. In some of our recent research exploring sensitivity to plural morphology during a listening-while-looking task, we found that 24-month-olds with higher vocabularies were sensitive to plural /s/ (as in cats; though not /z/, as in dogs; Davies, Xu Rattanasone, & Demuth, Reference Davies, Xu Rattanasone and Demuth2017). Further investigation confirmed that /s/ was longer in duration than /z/, and that this probably leads to its greater perceptual salience (despite the more frequent occurrence of /z/ as a plural morpheme). Thus, children who produce more words seem to be able to tune in to the greater perceptual salience of this /s/ allophone, attending to its meaning as a morphological unit and encoding it into their emerging grammar prior to doing the same with the voiced plural allomorph /z/.
Such findings suggest that there is a good correlation between the development of perceptual abilities and production, at least in global terms. However, identifying such correlations requires collecting multiple pieces of information from the same participants. For young infants, information on receptive vocabulary and/or the use of communicative gestures may provide a means for exploring such individual differences. Werker suggests that perceptual sensitivity develops as part of a multisensory system, where attention to multimodal (facial) cues may also play a role in learning how to articulate certain sounds. Perhaps children who attend to multimodal cues exhibit earlier onset of speech, and/or more targetlike use of early segments and prosodic word structures.
Recent research does suggest that motor and language development are closely linked (e.g., Iverson, Reference Iverson2010). This is further supported by research on infants fitted with cochlear implants, which finds that receptive communication abilities and fine motor skills before implantation, along with early gesture use, are significant predictors of vocabulary size at 12–15 months postimplantation (cf. Bavin, Sarant, Prendergast, & Leigh, Reference Bavin, Sarant, Prendergast and Leigh2017). This suggests that early motor abilities may correlate well with individual abilities in production, and perhaps perception as well.
SUMMARY
In sum, recent findings suggest that there may be a closer connection between perception and production abilities than often thought, with individual differences emerging in both. This can be illuminated by collecting multiple language/gesture/cognitive measures on the same children, shedding light on the nature of individual differences (cf. Cristia, Seidl, Junge, Soderstrom, & Hagoort, Reference Cristia, Seidl, Junge, Soderstrom and Hagoort2014). Such approaches are becoming more common in many other fields of psycholinguistic research (e.g., Montrul & Tanner, Reference Montrul and Tanner2017), and will hopefully become more common in infant/toddler research in years to come.
You have
Access
Werker provides a compelling story for how infants’ speech perceptual abilities develop and change over the first years of life. Less well understood is the relationship between perception and production abilities. Given pervasive evidence of individual differences in production, this raises many questions about possible individual differences in perception as well, and how this might influence the course of language acquisition more generally. This Commentary reviews some of the types of individual differences found in production, some of what we know about how these are linked to perception, and areas for further research.
It has long been known that children follow certain general patterns when acquiring their first words. However, there are also individual differences, in the timing of when children begin to produce their first words as well as in the actual phonological shape that these early words take. Thus, many children use their first words by 18 months, with a word spurt around 50 words, reaching 300 words by 24 months (cf., Fenson, Dale, Reznick, Bates, & Thal, Reference Fenson, Dale, Reznick, Bates and Thal1994). However, others start to use their first words as early as 10 or 11 months, and are using short sentences by 18 months (e.g., two of the children in the Providence Corpus; Demuth, Culbertson, & Alter, Reference Demuth, Culbertson and Alter2006). Others (“late talkers”) may only begin to use their first words after the age of 2, catching up around the age of 3 (cf., Rescorla, Roberts, & Dahlsgaard, Reference Rescorla, Roberts and Dahlsgaard1997). All the above is considered to be within the normal range for children with no obvious evidence of language disorders. Yet little is known about the possible differences in the timing or characteristics of children’s perceptual abilities, and if or how this might forecast these differences in production. Thus, although there are reports of perceptual sensitive at 6 months correlating with global measures of vocabulary size 1 or 2 years later (e.g., Tsao, Liu, & Kuhl, Reference Tsao, Liu and Kuhl2004), very little is known about the relationship between the perception and production of the same contrasts. For example, despite the fact that many children’s first words take the form of simple, “unmarked” segmental, syllable, and prosodic word structures (e.g., Demuth, Reference Demuth1995; Gnanadesikan, Reference Gnanadesikan2004), we know little about why this evolves differently depending on the child, even for a given language.
Thus, the field of child phonology, including children’s early word productions, appears to differ from that of the field of infant speech perception. However, this may be an artifact of the methods used. Studies of children’s phonological development have often taken the form of case studies, examining the speech productions of a few children, often longitudinally, providing an in-depth look at individual developmental patterns over time (e.g., Brown, Reference Brown1973). In contrast, infant speech perception studies have tended to involve groups of infants, reporting the development of language-specific perception abilities at a certain age. Although some theories of perception explicitly appeal to articulatory gestures in making claims about perception (e.g., Best’s, Reference Best1995, perceptual assimilation model), there has been little explicit research exploring the relationship between perception and production abilities in the same children. This is partly due to the different ages typically investigated, with much of the field of infant speech perception focusing on infants below the age of 2, and much of the child phonology literature focusing on children from the age of 2 onward. This, plus the different nature of the online versus speech production methods typically used, which necessitate the study of groups versus individuals, respectively, has led to little cross-talk between the two fields. Fortunately, this situation is slowly beginning to change, with concomitant advances in our theoretical understanding of the important connections between early perception and production.
In the following, I review some of what is known about individual differences in production, at the segmental level, interactions between segments and syllables, and at the level of prosodic words and prosodic morphology. I then briefly discuss recent research from our lab that begins to explore individual differences at the perception/production interface, discuss the possible mechanisms behind these individual differences, and identify areas for further research.
INDIVIDUAL DIFFERENCES IN THE PRODUCTION OF SEGMENTS, SYLLABLES, AND PROSODIC WORDS
Jakobson (Reference Jakobson1941) was one of the first to suggest that the segments children first use would be those that maximized contrasts. Thus, he proposed that the low vowel /a/ might be the first vowel to appear, along with the obstruent sonorant contrast /p/ versus /m/. Although the validity of Jakobson’s proposal has never been confirmed, many of these or similar segments do seem to be some of the first to appear in children’s early babbling and words, with consonants that are harder to produce, such as fricatives, affricates, and liquids later acquired (cf. Smit, Reference Smit1993). Thus, children’s early productions tend to include the more “unmarked” segments that occur widely across many languages. Of course, different languages have different phonological systems, and this also influences when and how phonemic contrasts are learned. For example, recent research on the early acquisition of Fulfulde (a West-Atlantic Niger–Congo language spoken in Mali and widely throughout West Africa) shows that, before 11 months, at least some children are using complex implosives and prenasal stops, and that implosives (e.g., /ɓ/) account for more than 37% of the obstruents one child produced (cf. Cissé, Demolin, & Vallée, Reference Cissé, Demolin and Vallée2011). It is not yet known if this was a particularly precocious child, or if this is a general pattern found for many Fulfulde-speaking children. It would then be interesting to know how and when perceptual sensitivity to, for example, implosive–nonimplosive contrasts emerges for such children.
Other more “marked” segments can also appear early in children’s speech. Dental clicks are produced early in languages like Xhosa (Gxilishe, Reference Gxilishe2004), whereas palatal alveolar clicks only appear later, and in Sesotho-speaking children’s words around the age of 3 (Demuth, Reference Demuth2003). One might expect perceptual sensitivity to all such consonants to appear long before the articulatory demands of production are met. The same might be thought to be true for segments like fricatives and liquids, where the articulatory challenges of producing such forms delays their appearance in children’s speech (cf. Lin & Demuth, Reference Lin and Demuth2015, for ultrasound study of English “l” development from 3–7 years).
Individual differences have long been documented at the segmental level in production, with some children preferring certain segments over others (Schwartz & Leaonard, Reference Schwartz and Leonard1982). For example, some have been shown to produce (many) words with labial–coronal sequences, or labial first words (cf. Yavaʂ, Reference Yavaʂ1995), depending on the language. Word truncation processes can also vary from child to child, where different segments may be preserved. When truncating trisyllabic words like banana to a disyllable, many children omit the first syllable altogether, producing a form like nana. Others, however, use the form bana, preserving the less sonorant (less marked) consonant in the onset position (cf. Demuth, Reference Demuth1996a; Pater, Reference Pater1997). Markedness in the selection of coda consonants also shows individual differences, with some English-speaking children’s first coda consonants being composed of less marked (more sonorant) codas (such as fricatives), whereas many others first use the most frequent (but more marked) coda consonant /t/ (Stites, Demuth, & Kirk, Reference Stites, Demuth and Kirk2004).
There are also many individual differences in the use of syllable structures and how these are produced. Some children begin using coda consonants with their first words, but many do not, producing predominantly reduplicated or consonant–vowel–consonant–vowel (CVCV) word forms for their first words (Demuth, Reference Demuth1995). A few, apparently precocious, children then go through a brief early stage of development (around 16 months) where they produce heavily released final (coda) consonants that can result in an “epenthetic” vowel, especially following voiced target codas (cf. Demuth et al., Reference Demuth, Culbertson and Alter2006; Vihman & Velleman Reference Vihman and Velleman2000). With respect to complex onsets, some children truncate word initial s-clusters by consistently dropping the /s/ (e.g., stop~top, slip~lip), whereas others preserve the more sonorant consonant, resulting in sop for stop, but lip for slip (cf. papers in Goad & Rose, Reference Goad and Rose2003). Individual differences in the use of consonant clusters are also found in Dutch-speaking children’s early words, with some showing earlier use of clusters in the onset (CCVC), but others first showing complexity in the coda (CVCC; Levelt, Schiller, & Levelt, Reference Levelt, Schiller and Levelt2000). Thus, there are many well-attested individual differences in both the timing and the realization of segments between children learning the same language. It is not clear if or how this might relate to perceptual differences as well. Little is known about individual differences in the perceptual sensitivity to such segmental/syllabic contrasts, and how this might relate to these individual differences in production.
Individual differences are also found in the form of children’s early prosodic words. For example, some Dutch-speaking children shift the placement of stress on words with a weak–strong (unstressed–stressed, iambic) stress pattern to a strong–weak (stressed–unstressed, trochaic) stress pattern, in keeping with the dominant stress pattern of the language (cf. Demuth, Reference Demuth1996b; Fikkert, Reference Fikkert1994). Although there has been much research on infant sensitivity to language-specific stress patterns (e.g., Jusczyk, Cutler, & Redanz, Reference Jusczyk, Cutler and Redanz1993; Yeung & Nazzi, Reference Yeung and Nazzi2014), little is known about individual differences in perception, and how this may relate to these differences in production.
Some children also tend to use “proto function words” often known as “fillers” (e.g., a wan a more “I want some more”), whereas others do not (cf. Peters, Reference Peters1983). It is unclear why this is the case, and if children’s perceptual sensitivities to function words (e.g., Shi, Werker, & Cutler, Reference Shi, Werker and Cutler2006) show concomitant individual differences. Some Spanish-speaking children actively truncate lexical items even while producing early “proto-articles” (amueka “la muñeca” (the doll)), while others never truncate lexical forms, only producing articles with fully formed words (cf. Demuth, Patrolia, Song, & Masapollo, Reference Demuth, Patrolia, Song and Masapollo2012). Individual differences are also found in English-speaking children’s use of articles, with most children prosodifying articles with the proceeding monosyllabic word (e.g., He [likes the] dolly; Gerken, Reference Gerken1996), but some actually stressing the article, initially producing it as its own prosodic word around the age of 20 months (He [likes] [the] dolly; cf. Demuth & McCullough, Reference Demuth and McCullough2009). Again, it is likely that perception of such forms is robust (cf. Gerken & McIntosh, Reference Gerken and McIntosh1993), but some of our resent research below suggests a close connection between perception and production.
INDIVIDUAL DIFFERENCES AT THE PERCEPTION/PRODUCTION INTERFACE
It is unclear if any of the individual differences in production discussed above correlate with differences in perception. However, there are indications that some of them might. For example, we have found that, in spontaneous speech, 2-year-olds tend to produce more inflectional morphemes, such as 3rd person singular –s, when these occur utterance finally compared to utterance medially (Song, Sundara, & Demuth, Reference Song, Sundara and Demuth2009). The final syllables of a phrase undergo phrase-final lengthening (Klatt, Reference Klatt1976; Lehiste, Reference Lehiste1973); we thus thought that phrase-final contexts might provided a longer durational context for all segments (and morphemes) to be produced. However, this phrase-final lengthening might also enhance the perceptual salience of these morphemes, making them easier to perceive as well. This was tested and found to be the case: both 20- and 27-month-olds showed sensitivity to 3rd person singular inflectional morphemes utterance finally, but not utterance medially, suggesting a close connection between perception and production (Sundara, Demuth, & Kuhl, Reference Sundara, Demuth and Kuhl2011). This finding also suggests that higher perceptual salience may facilitate “learning,” and this will then also be reflected earlier in production.
However, most verbs in English occur utterance medially, raising questions about when and how children begin to perceive and produce inflected verbs in this position. Further investigation revealed that those infants who showed higher 3rd person singular production abilities in an elicited imitation task also showed greater sensitivity to the presence versus (ungrammatical) absence of the morpheme utterance medially (e.g., She cries/cry now) during a listening task (Sundara et al., Reference Sundara, Demuth and Kuhl2011). Such findings suggest a tight correlation between perception and production at the level of the individual.
What could be the mechanism underlying these individual differences? Werker suggests that perceptual sensitivity interacts with meaning and referential context. We know that individual children’s receptive and expressive vocabulary sizes show large individual differences from the time these can be measured. Perhaps this is a result of more efficient perceptual development, leading to earlier production as well? Recent findings suggest that this may be the case. In some of our recent research exploring sensitivity to plural morphology during a listening-while-looking task, we found that 24-month-olds with higher vocabularies were sensitive to plural /s/ (as in cats; though not /z/, as in dogs; Davies, Xu Rattanasone, & Demuth, Reference Davies, Xu Rattanasone and Demuth2017). Further investigation confirmed that /s/ was longer in duration than /z/, and that this probably leads to its greater perceptual salience (despite the more frequent occurrence of /z/ as a plural morpheme). Thus, children who produce more words seem to be able to tune in to the greater perceptual salience of this /s/ allophone, attending to its meaning as a morphological unit and encoding it into their emerging grammar prior to doing the same with the voiced plural allomorph /z/.
Such findings suggest that there is a good correlation between the development of perceptual abilities and production, at least in global terms. However, identifying such correlations requires collecting multiple pieces of information from the same participants. For young infants, information on receptive vocabulary and/or the use of communicative gestures may provide a means for exploring such individual differences. Werker suggests that perceptual sensitivity develops as part of a multisensory system, where attention to multimodal (facial) cues may also play a role in learning how to articulate certain sounds. Perhaps children who attend to multimodal cues exhibit earlier onset of speech, and/or more targetlike use of early segments and prosodic word structures.
Recent research does suggest that motor and language development are closely linked (e.g., Iverson, Reference Iverson2010). This is further supported by research on infants fitted with cochlear implants, which finds that receptive communication abilities and fine motor skills before implantation, along with early gesture use, are significant predictors of vocabulary size at 12–15 months postimplantation (cf. Bavin, Sarant, Prendergast, & Leigh, Reference Bavin, Sarant, Prendergast and Leigh2017). This suggests that early motor abilities may correlate well with individual abilities in production, and perhaps perception as well.
SUMMARY
In sum, recent findings suggest that there may be a closer connection between perception and production abilities than often thought, with individual differences emerging in both. This can be illuminated by collecting multiple language/gesture/cognitive measures on the same children, shedding light on the nature of individual differences (cf. Cristia, Seidl, Junge, Soderstrom, & Hagoort, Reference Cristia, Seidl, Junge, Soderstrom and Hagoort2014). Such approaches are becoming more common in many other fields of psycholinguistic research (e.g., Montrul & Tanner, Reference Montrul and Tanner2017), and will hopefully become more common in infant/toddler research in years to come.
ACKNOWLEDGMENTS
This research was supported by Macquarie University, the Australian Research Council (ARC) Centre of Excellence in Cognition and Its Disorders Grant CE110001021 (awarded to Crain et al.), and ARC Grant FL130100014 (to K.D.).