Abstract
Although several studies initially supported the proposal by Nespor et al. (Nespor, Marina, Marcela Peña & Jacques Mehler. 2003. On the different roles of vowels and consonants in speech processing and language acquisition. Lingue e Linguaggio 2. 221–247) that consonants are more informative than vowels in lexical processing, a more complex picture has emerged from recent research. Current evidence suggests that infants initially show a vowel bias in lexical processing and later transition to a consonant bias, possibly depending on the characteristics of the ambient language. Danish infants have shown a vowel bias in word learning at 20 months—an age at which infants learning French or Italian no longer show a vowel bias but rather a consonant bias, and infants learning English show no bias. The present study tested whether Danish 20-month-olds also have a vowel bias when recognizing familiar words. Specifically, using the Intermodal Preferential Looking paradigm, we tested whether Danish infants were more likely to ignore or accept consonant than vowel mispronunciations when matching familiar words with pictures. The infants successfully matched correctly pronounced familiar words with pictures but showed no vowel or consonant bias when matching mispronounced words with pictures. The lack of a bias for Danish vowels or consonants in familiar word recognition adds to evidence that lexical processing biases are language-specific and may additionally depend on developmental age and perhaps task difficulty.
1 Introduction
Speech sounds used in the world’s languages are usually described in two broad categories, namely vowels and consonants (Ladefoged and Maddieson 1996). According to a line of research starting with Nespor et al. (2003), consonants are more informative than vowels for lexical processing. Research findings in the decade after the proposal by Nespor et al. generally concurred on a greater weight given to consonants than vowels in lexical processing in both children and adults in a range of languages. However, recent research has drawn a more complex picture, finding a vowel bias in certain languages in children or adults or a lack of bias in children or adults. In addition, a developmental pattern was suggested by Nazzi et al. (2016), such that languages universally favor an initial vowel bias in lexical processing, with a later transition to a consonant bias depending on the specific phonological and/or lexical characteristics of each language. Danish was the first language for which a relatively late vowel bias in infancy was documented (20 months of age) for learning of novel words. The present study aimed to lay one more piece in the puzzle concerning use of vocalic and consonantal information in lexical processing by examining whether a vowel bias (or consonant bias) exists in Danish-learning 20-month-olds’ processing of familiar words.
1.1 Vowels and consonants in lexical processing
Language is encoded in sequences of speech sounds. Acoustically, vowels are a relatively homogenous group of speech sounds. They are typically voiced and vary primarily in the spectral domain, namely in the pattern of formant frequencies (shaped by the configuration of the articulators) although a range of secondary features also distinguish vowels in some languages (Ladefoged and Maddieson 1990). Consonants, on the other hand, form a quite heterogenous group of speech sounds. Stop consonants like [b] or [k] are relatively brief sounds, often consisting of a burst and formant transitions into or out of neighboring sounds. Fricatives like [f] or [z] are longer, relatively steady-state noises. Both stops and fricatives may be voiced or voiceless. Other steady-state consonants like nasals or lateral approximants are typically voiced and share some characteristics with vowels, whereas central approximants like [j] and [w] have very vowel-like acoustic characteristics.
Both vowels and consonants are used to contrast meaning in all of the world’s languages. Some languages additionally contrast meaning by different tone heights or tone movements, such as Mandarin and Cantonese, consonant length, such Italian or Finnish, or vowel length, such as Finnish or Danish. Danish additionally has phonological laryngilization or creaky voice (so-called “stød”), which may occur in long vowels or short vowel + sonorant sequences. The laryngilization of “Stød” typically co-occurs with an abrupt dip in fundamental frequency; therefore “stød” may also be viewed as phonological tone change (Grønnum et al. 2013). Interestingly, Mandarin tones, especially tone 3, are often associated with creaky voice (Kuang 2017).
In sum, consonants are a larger and acoustically more diverse group of sounds than vowels are, and, perhaps for these reasons, consonants outnumber vowels in most of the world’s languages (Nespor et al. 2003).
The heterogeneity of consonants and their higher frequency of use to contrast meaning in the world’s languages led Nespor et al. (2003) to hypothesize that consonantal information is more informative for lexical processing than vocalic information is, and that this consonantal bias is universal across languages. Nespor et al. (2003: 206) proposed an additional advantage for consonants, namely that “they tend to disharmonize within a word, i.e., to become more distinctive”. Vowels, on the other hand, do not tend to disharmonize, but rather to harmonize throughout a domain, with a consequential loss of distinctive power. Nespor et al. (2003) proposed that this additional consonantal advantage for encoding meaning may explain a consonantal lexical processing bias also in languages with a balanced number of vowels and consonants.
The hypothesis that consonants are universally more important than vowels for lexical processing initially received support from studies of multiple languages in both speech perception (e.g., Bonatti et al. 2005; Cutler et al. 2000; van Ooijen 1996), written word perception (Acha and Perea 2010; New et al. 2008) and neurological research (Caramazza et al. 2000; Carreiras and Price 2008; Carreiras et al. 2009). However, several more recent studies from diverse languages have drawn a more complex picture.
1.2 Vowels and consonants in infants’ lexical processing
The review of lexical processing biases provided by Nazzi et al. (2016) suggest a developmental transition from an initial vowel bias to an eventual consonant bias. However, the timing of the transition—if happening at all—varies across languages. The following is a review of representative studies documenting the existence of biases and bias transitions in lexical processing in several languages, beginning with languages with an early transition and ending with languages with a late transition or no documented transition. Figure 1 gives an overview of the findings of biases according to language, approximate age, and whether the findings pertain to familiar word processing or word learning. Presently, only results for (monolingual) native speakers are discussed.
Overview of studies reporting lexical processing bias towards consonants, vowels, or no bias as a function of language, approximate age (rounded to nearest year for infants and toddlers), and type of task (word learning or processing of familiar words). The figure is inspired by and updates Figure 1 of Nazzi et al. (2016). Note: 1 (Havy and Nazzi 2009); 2 (Nazzi 2005); 3 (Nazzi and Polka 2018); 4 (Havy et al. 2014); 5 (Bouchon et al. 2015); 6 (Von Holzen and Nazzi 2020); 7 (Poltrock and Nazzi 2015); 8 (Delle Luche et al. 2014); 9 (Benavides-Varela et al. 2012); 10 (Hochman et al. 2018); 11 (Hochmann et al. 2011); 12 (Bonatti et al. 2005); 13 (Floccia et al. 2014); 14 (Nazzi et al. 2009); 15 (Creel et al. 2006); 16 (Delle Luche et al. 2017); 17 (Mani and Plunkett 2007); 18 (Mani and Plunkett 2010); 19 (Ratnage et al. 2023); 20 (van Ooijen 1996); 21 (Højen and Nazzi 2016); 22 (present study); 23 (Højen et al. in preparation); 24 (Chen et al. 2021); 25 (Gómez et al. 2018); 26 (Poltrock et al. 2018); 27 (Wewalaarachchi et al. 2017); 28 (Wiener and Turnbull 2016); 29 (Wiener 2020).
That French-learning infants have an initial vowel bias was documented by Bouchon et al. (2015), who found that 5-month-olds’ recognition of their own name was impaired by vowel changes but not consonant changes. Von Holzen and Nazzi (2020) in a similar experiment on French-learning 5-, 8-, and 11-month-olds replicated the vowel bias for 5- and 8-month-olds but found a consonant bias in the group of 11-month-olds. This suggested that the vowel to consonant bias transition happens between 8 and 11 months in French-learning infants. Multiple studies had previously documented a consonant bias in word learning and/or recognition in older French-learning children or adults (Delle Luche et al. 2014; Havy et al. 2014; Havy and Nazzi 2009; Nazzi 2005; Nazzi and Polka 2018; Nazzi et al. 2009; Poltrock and Nazzi 2015). Importantly, Nazzi and Polka (2018) showed that, in a word learning task, Canadian French learning 20-month-olds showed a consonant bias even in words with an initial vowel. That is, this study resolved a confound between segment type (vowel or consonant) and position in the word in previous studies; French-learning infants’ consonant bias in lexical processing does not depend on position in the word.
In Italian-learning infants, an early transition in bias similar to that in French has been documented. In a study of newborn babies’ ability to memorize newly learned words, Benavides-Varela et al. (2012) using infrared spectroscopy, found that newborns were better able to use vowel than consonantal information. A study of 6–7-month-old Italian infants’ ability to learn new words and pair them with objects showed that, at this age, Italian infants still had a vowel bias; they relied more on vowel than consonant information when associating familiarized words—with a consonant or a vowel change—with objects (Hochmann et al. 2018). However, by the age of 12 months, Italian infants were found to favor consonant information in a study using the same paradigm (Hochmann et al. 2011). That is, by age 1 Italian infants show a consonant bias. The consonant bias seems to be retained in word learning in adulthood, although the experiment showing the consonant bias in Italian adults used concatenated syllables, which may sound somewhat different than natural speech (Bonatti et al. 2005).
For early acquisition of English, the pattern of early lexical processing biases differs from that of French and Italian. Delle Luche et al. (2017) using the paradigm of Bouchon et al. (2015), did not find either a vowel or a consonant processing bias in English-learning 5-month-olds’ recognition of their own name in vowel- or consonant-mispronounced rendition. A lack of significant difference does not imply similarity. However, the lack of a lexical processing bias towards vowels or consonants has also been found in English-learning infants’ word recognition using the Head-turn Preference Procedure at the age of 11 months (Ratnage et al. 2023) and—using the Intermodal Preferential Looking paradigm (IPL)—at the age of 12 months (Mani and Plunkett 2010) as well as 15, 18, and 24 months (although latency measures of gaze shift indicated a consonant bias in 15-month-olds, which looking time did not). In addition, studies of word learning failed to find a processing bias in English at the age of 16 or 23 months (Floccia et al. 2014). Taken together, these studies suggest a symmetry rather than a consonant or vowel bias between 5 and 24 months of age in English-learning infants; but at the age of 30 months, a shift towards the more typical consonant bias was documented for English-learning infants using a name-based categorization task (Nazzi et al. 2009). In this task, vowel and consonant mispronunciations were pitted directly against each other, by presenting a word that could be matched to either of two targets, depending on whether the infant chose to ignore a mispronounced vowel or a mispronounced consonant (e.g., matching [pide] with either of the two targets [tide] or [pyde]).
For Danish, there are no studies documenting lexical processing biases in the first year of life. But one study reported a vowel bias in Danish-learning 20-month-olds in an interactive word-learning task (Højen and Nazzi 2016). The infants could learn separate names for new objects if they differed by a single vowel (e.g., sil–sel), but not if they differed by a single consonant (e.g., bem–pem). However, at some yet undetermined point in development, native speakers of Danish develop a consonant bias in lexical processing. Recently a word reconstruction study—using the paradigm of van Ooijen (1996)—found that Danish adults have a consonant bias in recognition of mispronounced words (Højen et al. in preparation). Specifically, native Danish adults made significantly more errors when required to change a consonant to reconstruct a mispronounced word (e.g., reconstruct kebra to zebra) than when required to change a vowel (e.g., kebra to cobra) or when having free choice (e.g., kebra to either cobra or zebra). Response times were slightly, but not significantly, faster in the vowel change than the consonant change condition. This pattern of results is an exact replication of those for English (van Ooijen 1996), Spanish, and Dutch (Cutler et al. 2000).
For Cantonese, a tone language, a vowel bias was found by Chen et al. (2021) in 30-month-olds. The authors used eye-tracking to examine learning of words that differed minimally by either a tone, a consonant, or a vowel. Whereas proportional looking time measures did not suggest word learning in the 20-month-olds, the 30-month-olds showed learning of words that differed by a vowel, but not a consonant. This finding suggests a vowel bias in 30-month-olds’ word learning in Cantonese. A study of adult Cantonese native speakers’ word learning, using the same eye-tracking paradigm as that later used by Chen et al. (2021), revealed no vowel bias or consonant bias (Poltrock et al. 2018). This may suggest that the Cantonese vowel bias disappears somewhere between 30 months and adulthood. However, Gómez et al. (2018) in a study of segmentation of concatenated nonwords, found a vowel bias in native Cantonese adults. The participants were able to use statistical cues for segmentation only if those cues were carried by vowels. However, as noted earlier regarding the Bonatti et al. (2005), concatenated syllables may sound somewhat different than natural speech.
For Mandarin, which is also a tone language, Wewalaarachchi et al. (2017) found a vowel bias in 24-month-olds’ recognition of familiar words. The infants rejected mispronounced words whether mispronounced in a vowel or a consonant, but were non-significantly more accepting of consonant mispronunciations. However, time course analyses revealed significant processing differences, indicating a significant vowel bias. For adults’ processing of familiar words in Mandarin, Wiener and Turnbull (2016) using the van Ooijen word reconstruction paradigm described earlier, found a vowel bias rather than a consonant bias. The study involved both tone, vowel and consonant mispronunciations. The participants were most accurate and fastest when changing tones, but least accurate and slowest when changing vowels, suggesting a vowel bias in Mandarin lexical processing. The result was replicated by Wiener (2020).
In summary, lexical processing biases vary across languages and across developmental age. The current evidence suggests that infants begin language acquisition relying or focusing on vowels more than consonants but transition from this early vowel bias to a consonant bias at some point in development that varies across languages, possibly depending on the characteristics of the language (for a review, see Nazzi and Cutler 2019). This may suggest that perceptually focusing on vowel information is easier and may bootstrap word learning, but that consonants are more informative in the long run for establishing lexical distinctions for many languages, causing a shift in processing bias in favor of consonants once the perceptual system gains more robustness. This is in line with the hypothesis that infants selectively focus attention on certain aspects of the acoustic signal when processing demands are too high (Fennell and Werker 2003; Stager and Werker 1997).
1.3 The present study
As noted in the foregoing, and as shown in Figure 1, the same processing bias (vowel, consonant or no bias) has been found for word learning and familiar word recognition whenever those two types of tasks have been examined for the same age in the same language. However, this comparison is only possible to make for French and English 1- or 2-year-olds as well as English adults.
The question addressed by the following experiment was whether Danish 20-month-olds show a vowel bias in familiar word recognition, that is, the same bias as previously found in 20-month-olds’ Danish word learning, or a consonant bias such as found in native Danish adults’ familiar word recognition. As shown in Figure 1, we have foreshadowed the results of the experiment reported below, which showed a lack of processing bias toward consonant or vowel information.
The experiment examined processing bias in familiar Danish words by testing the use of vocalic and consonantal information in the perception of Danish consonant-vowel-consonant (CVC) words using the IPL procedure. The following explanation uses English words for illustration. Two pictures were shown at the same time on each side of a screen, for example a book and a sock. A name was presented auditorily either correctly pronounced (book [bʊk] or sock [sɑk]) or incorrectly pronounced (bock [bɑk] or sook [sʊk]), and the effect of naming on infants’ looking preference for the pictures was determined. When mispronounced, the name presented a conflict task (Nazzi et al. 2009), with the possibility of hearing the mispronounced name as either of the words illustrated by the pictures, if the infant ignored either a consonant or a vowel mispronunciation and changed it to the correct sound. Thus, the infant could identify bock [bɑk] as book if ignoring the vowel mismatch, or identify bock as sock if ignoring the consonant mismatch. Vice versa, sook [sʊk] could be identified as book if ignoring the consonant mismatch, or as sock if ignoring the vowel mismatch. In this way, each pair of mispronounced words served as a balancing control to counter potential biases with respect to word familiarity or phonetic distance between the target sound and mispronounced sounds.
As in the study by Nazzi et al. (2009), this experiment examined the possible confound effect between V/C mispronunciation and position in the word by having two conditions. In the C-first condition, the mispronounced word could be heard as a familiar word by ignoring mispronunciation of the first consonant in a CVC word, or heard as another familiar word by ignoring mispronunciation of the vowel. In the C-last condition, familiar words could be heard by ignoring mispronunciation of the vowel or the last consonant in CVC words.
2 Methods
2.1 Participants
Thirty-two healthy infants from monolingual Danish-speaking homes participated. They were recruited via e-mails to parents of children living in or near the city of Odense, Denmark, who had children around 20 months of age (based on national registry information available to researchers). The ages ranged from 18 months, 28 days to 21 months, 14 days (mean = 20;4 months). The parents reported no history of ear infection in the children. Sixteen infants were randomly assigned to each condition (C-first/C-last). An additional seven infants were tested but were excluded because of experimenter error (four) or because of failure to reach task engagement criterion (three; total looking time at either picture averaging less than 1,000 ms during the 1,510 ms observation window, see procedure). For each infant, a Danish MacArthur-Bates communicative development inventory (CDI) form was completed (Bleses et al. 2008; Fenson et al. 2007) in order to examine the relation of vocabulary size with word identification in the experiment. The CDI is a standardized and validated measure of children’s vocabulary skills based on parent report. For each of 725 words, parents indicate whether the child produces and/or comprehends the word. Parents completed the report prior to visiting for the experiment.
2.2 Auditory stimuli
The auditory stimuli consisted of correctly pronounced (CP) CVC words, expected to be familiar, and mispronounced (MP) word stimuli that differed from the CP-words either in the initial consonant or the vowel (C-first condition) or in the final consonant or the vowel (C-last condition), see Table 1. There were four correctly pronounced words in each condition; two of the words were used in both conditions.
Eight familiar Danish words with correct pronunciation (CP) or mispronounced (MP). Two conditions: C-first (consonant-before-vowel mispronunciation) and C-last (vowel-before-consonant mispronunciation). The number of phonological features changed in the mispronunciation is given after each mispronounced word.
| Word CP | English | C MP | V MP | |
|---|---|---|---|---|
| C-first | sut | ‘pacifier’ | kut (2) | sat (3) |
| kat | ‘cat’ | sat (2) | kut (3) | |
| bil | ‘car’ | fil (2) | bul (2) | |
| ful | ‘bird’ | bul (2) | fil (2) | |
| C-last | bil | ‘car’ | big (2) | bol (3) |
| bog | ‘book’ | bol (2) | big (3) | |
| kat | ‘cat’ | kap (1) | kot (3) | |
| kop | ‘cup’ | kot (1) | kap (3) |
-
Notes: For ease of reading the table, the word fugl (with a silent g) is misspelled here as ful. Broad phonetic transcription of each CP-word: [sut kæt biːʔl fuːʔl bɔːʔw kʌp]. Non-mispronounced consonant and vowel qualities were retained in MP-words.
The stimuli were produced by a young native Danish female in a moderately child-directed voice and recorded in a sound-treated studio onto a Fostex harddisk recorder (22.05 kHz, 16 bit) using a large diaphragm Audio-Technica microphone placed circa 20 cm from the mouth behind a pop filter, resulting in high-quality audio files. The words were produced in the carrier sentence, Se, en _____! (“Look, a _____”). First, the correctly pronounced words for the CP-first condition were recorded, then the mispronounced words. Subsequently, the correctly pronounced words and then the mispronounced words were recorded for the CP-last condition.
The stimulus words including the preceding determiner (en) were edited out, normalized for rms amplitude and spliced onto copies of a single carrier sentence (in effect just the word Se). Each resulting stimulus sentence was saved in a separate wav-file.
Each of the eight MP-stimuli could be the mispronounced form of two of the eight CP-words. Whether it was heard as one or the other of the two, depended on whether the vowel mispronunciation or the consonant mispronunciation was ignored/changed. For example, the mispronounced form kut could be turned into the correct word kat by changing the vowel or turned into the correct word sut by changing the initial consonant (see example in Figure 1). Opposite relations held between the MP-form sat and the two CP-words kat and sut: Sat could be turned into the word sut by changing the vowel or into kat by changing the initial consonant.
The number of phonological feature changes in vowel and consonant mispronunciations are shown in parentheses in Table 1. The vowel mispronunciations involved more phonological feature changes than the consonant mispronunciations in six out of eight cases. This imbalance could result in a bias towards being more accepting of consonant mispronunciations, if the feature imbalance translates into an imbalance in perceptual saliency. However, the results of Nazzi et al. (2009, see introduction) indicate that there is far from a one-to-one relation between feature differences and perceptual saliency for infants.
Perceptual saliency for infants may be related to acoustic differences more likely than to phonological feature differences. Table 2 shows acoustic characteristics of each CP word and MP word in each of the two conditions. Inspection of the mean values in the CP and MP words in each condition indicates little difference as a function of pronunciation. A single exception is the mean F3 value in the C-last condition, which was about 200 Hz lower in the MP than in the CP condition. This difference was driven solely by a considerably lower F3 in the [i] of the MP word big than the CP word bil, likely due to coarticulation with the final [w] in the MP word big.
Word duration and segment duration of the first consonant (C1) the vowel (V) and the last consonant (C2) for each of the CP-words and MP-words, in addition to F0 and the first three formant frequencies (F1, F2, F3) taken at vowel midpoint for words used in the C-first and the C-last condition.
| C-first condition | Word dur. ms | C1 dur. ms | V dur. ms | C2 dur. ms | V F0 Hz | V F1 Hz | V F2 Hz | V F3 Hz | |
|---|---|---|---|---|---|---|---|---|---|
| Correct pronunc. | sut | 401 | 156 | 84 | 161 | 267 | 314 | 1,276 | 2,625 |
| kat | 297 | 62 | 99 | 136 | 243 | 723 | 2,042 | 2,955 | |
| bil | 282 | 10 | 180 | 92 | 296 | 318 | 2,680 | 3,681 | |
| ful | 446 | 148 | 173 | 125 | 308 | 330 | 924 | 2,792 |
|
| Mean | 357 | 94 | 134 | 129 | 279 | 422 | 1,731 | 3,013 |
|
| Mispronounced | kut | 309 | 83 | 78 | 148 | 288 | 321 | 1,028 | 2,808 |
| sat | 347 | 123 | 111 | 113 | 245 | 712 | 1,954 | 2,989 | |
| fil | 470 | 143 | 191 | 136 | 304 | 326 | 2,734 | 3,332 | |
| bul | 387 | 12 | 190 | 185 | 307 | 319 | 928 | 2,752 |
|
| Mean | 367 | 92 | 138 | 136 | 282 | 421 | 1,700 | 2,994 | |
| C-last condition | Word dur. ms | C1 dur. ms | V dur. ms | C2 dur. ms | V F0 Hz | V F1 Hz | V F2 Hz | V F3 Hz | |
|---|---|---|---|---|---|---|---|---|---|
| Correct pronunc. | bil | 282 | 10 | 180 | 92 | 296 | 318 | 2,680 | 3,681 |
| bog | 319 | 10 | 171 | 138 | 268 | 497 | 1,064 | 2,634 | |
| kat | 297 | 62 | 99 | 136 | 243 | 723 | 2,042 | 2,955 | |
| kop | 373 | 55 | 115 | 203 | 248 | 718 | 1,260 | 2,711 |
|
| Mean | 318 | 34 | 141 | 142 | 264 | 564 | 1,762 | 2,995 |
|
| Mispronounced | big | 306 | 10 | 146 | 150 | 279 | 291 | 2,488 | 2,801 |
| bol | 309 | 11 | 197 | 101 | 275 | 508 | 1,100 | 2,684 | |
| kap | 341 | 63 | 123 | 155 | 249 | 908 | 1,974 | 2,996 | |
| kot | 305 | 69 | 114 | 122 | 241 | 693 | 1,290 | 2,741 |
|
| Mean | 315 | 38 | 145 | 132 | 261 | 600 | 1,713 | 2,806 | |
2.3 Visual stimuli
The visual stimuli were cartoon drawings of the objects named by the CP-words. The pictures (width: 32 cm, height: 18 cm) were shown with a distance of 75 cm between them in the upper corners of a tv screen about 130 cm from the child. An example of the stimuli is shown in Figure 2.
Illustration of test setup seen from the infant’s perspective. The example is with an MP-trial with the child hearing kut. The child could identify kut with the picture of a “kat” by changing/ignoring the vowel, or identify kut with “sut” by changing/ignoring the first consonant.
2.4 Procedure
Before the experiment, each participating child’s parent confirmed that the four real words were known by the child. The experiment took place in a dim and visually neutral studio. The infant sat on their parent’s lap. Just underneath the monitor, two speakers played the stimuli. Just over the monitor three digital cameras recorded the infant’s face for later scoring of eye-movements.
CP-trials had a target and a distracter as in previous IPL research. However, MP-trials had no distracter but rather two possible targets, namely a vowel target (V-target) and consonant target (C-target). The V-target in MP-words had the correct vowel but the wrong initial/final consonant. The C-target had the correct initial consonants but the wrong vowel.
The existence of two possible targets in MP-trials required strict control over lexical priming effects from similar sounding words in preceding correct trials, which was ensured. If, for example, the mispronounced word sat were preceded more often by kat than sut, the infants might be primed more often to perceive sat as kat than sut.
Each infant was tested with a block of 16 trials. In each condition, the two picture pairs (“sut—kat” and “bil—fugl” or “bil—bog” and “kop—kat”) were each presented on four CP-trials and four MP-trials. On the four CP-trials for each picture pair, each picture was correctly named by the auditory stimulus on two trials, once on the left and once on the right side. Similarly, in MP-trials, each picture was the V-target or the C-target two times, once left and once right. To reduce priming effects, the two picture pairs were presented on every second trial; MP-words (e.g., kut) could never be immediately preceded by one of their two CP-forms (kat/sut).
The two CP-words that related to each MP-form came once each in a CP-trial two trials before the trial in which the MP-word occurred. For example, for the MP-form kut, the auditory stimulus two trials earlier was one time kat and one time sut. As a consequence of this precaution to control for priming effects, the blocks had to begin with two CP-trials and end with two MP-trials, and the trial structure throughout each block was CP-CP-MP-MP-CP-CP-MP-MP etc. For each child, the stimuli were balanced with respect to the left/right position of the target (including V- or C-targets) and the left/right position of the target (including V- or C-targets) of the preceding trial. The first picture pair was chosen randomly for each child with blocking within four subsequent children, such that all combinations of picture pair and auditory stimulus came equally frequently as the first trial.
Each trial lasted 5,000 ms, beginning with the display of two pictures and ending with display offset. The infant was given a couple of seconds to look at the pictures before audio was played. The auditory stimulus was timed such that the onset of the target word came after 2,250 ms, and the infant’s looking behavior was observed. The time window for observation did not begin immediately with the onset of the target word, but was delayed in order to give the infant time to shift gaze in response to the auditory stimulus. Previous research often delayed the onset of the observation window relative to target word onset by 367 ms (e.g., Mani and Plunkett 2007; Swingley and Aslin 2000). In the present study, the observation window was delayed slightly longer given that also word-final sounds were examined, and the infant should be given time to shift gaze also in response to the last sound. The observation window thus began 500 ms after word onset, and it lasted until 2,010 ms after word onset. The observation window of 1,510 ms was similar to that of some previous IPL studies (e.g., Swingley and Aslin 2002; Swingley and Aslin 2000, 1,633 ms) but slightly shorter than that of other studies (e.g., Mani and Plunkett 2008; Mani and Plunkett 2007, 2,133 ms).
During the observation window, the infant’s fixation of each picture was scored on a frame-by-frame basis from the digital video recording, and the number of frames with eye fixation on the target (or, in MP-trials, the vowel target) was divided by the total number of frames in the observation window to obtain the percent (vowel) target looks on each trial. Unlike some previous research, looking preferences in the pre-naming phase were not scored. A possible preference for certain pictures could not introduce a bias because each picture was the target, distractor, V-target or C-target the same number of times with a complete and balanced rotation of left/right position and target/distractor status in the trial.
Infants were first familiarized with each picture, which was presented and correctly named one at a time. Subsequently, the test began with a filler trial using non-test auditory and visual stimuli. A similar filler trial came for every four test trials to maintain interest.
3 Results
As in previous research, children were expected to show a preference for looking at the target when accepting an auditory stimulus as the target name. Percent looking time at the target in CP-trials or at the V-target in MP-trials in the observation window was calculated.
The results shown in Figure 3 indicate a preference for looking at the target in CP-trials and approximately chance-level preferences for the V-target versus C-target in MP-trials. The results were submitted to a 2 × 2 mixed-design ANOVA with condition (C-first, C-last) as a between-subjects factor, and pronunciation (CP, MP) as a within-subjects factor.
Mean percent looking at the target in CP-trials or the V-target in MP-trials for 32 20-month-old Danish infants; error bars: ±1 SE.
The results revealed a significant main effect of pronunciation (F(1, 30) = 24.72, p < 0.001, partial η 2 = 0.45), a nonsignificant main effect of condition (F(1, 30) = 0.07, p = 0.790, partial η 2 < 0.01), and a non-significant interaction (F(1, 30) = 0.10, p = 0.756, partial η 2 < 0.01). The infants were thus significantly more likely to look at the target in CP-trials than at the V-target (or C-target) in MP-trials.
To determine the statistical significance of a potential deviation from chance level, the results were submitted to two-tailed single-sample t-tests. For MP words the t-tests revealed non-significant differences from chance for the V-target in both the C-first condition (t(15) = −0.973, p = 0.346) and C-last condition (t(15) = 1.748, p = 0.101). In contrast, t-tests for CP-words revealed significantly above-chance preferences for the target in both the C-first condition (t(15) = 4.101, p < 0.001) and the C-last condition (t(15) = 5.341, p < 0.001).
Two exploratory analyses were carried out to examine whether phonetic or phonological differences between the word pairs in each condition influenced looking behavior. One might imagine, for example, that if the [b]–[f] distinction is less salient than the [s]–[k] distinction, the infants would be more likely to ignore [b]–[f] mispronunciations than [s]–[k] mispronunciations. Or if number of phonological features involved in mispronunciation translates into degree of perceptual salience, variance in number of phonological features might influence looking behavior in the two word pairs in each condition.
In the C-first condition, the word pairs required substitution of the vowels [u]/[æ] or [iːʔ]/[uːʔ] to match an MP stimulus to a target picture, or substitution of the consonants [s]/[k] or [b]/[f]. Vowel target looks in MP trials for the two word pairs were very similar, 49 % and 46 %. In the C-last condition, the word pairs required substitution of the vowels [iːʔ]/[ɔːʔ] or [æ]/[ʌ] to match an MP stimulus to a target picture, or substitution of the consonants [l]/[w] or [t]/[p]. Vowel target looks in MP trials for the two word pairs were also similar, 56 % and 53 %. Neither difference approached significance (ps > 0.529). The exploratory analyses thus suggest that neither phonetic nor phonological differences between mispronunciations across word pairs influenced looking behavior.
Finally, associations between CDI vocabulary scores (productive and receptive) and preference for the target in CP-trials or the V-target in MP-trials were examined in bivariate correlation analyses. Neither analysis revealed a significant correlation. This result aligns with most previous related research (e.g., Bailey and Plunkett 2002; Ballem and Plunkett 2005; Swingley and Aslin 2002), although Mani and Plunkett (2010) found an association between vocabulary score and sensitivity to vowel mispronunciations, but not consonant mispronunciations in infants.
4 Discussion and conclusion
One important result of the present study was the finding of recognition of familiar Danish CVC words by native Danish 20-month-olds when correctly pronounced, but a lack of recognition of words when mispronounced. This is evidence that native Danish 20-month-olds have a relatively accurate phonetic representation of familiar words, which has not been documented before.
Regarding lexical processing bias, no vowel or consonant bias in familiar word recognition was found when, in a conflict task, the infants had to ignore a vowel or a consonant mispronunciation to match a mispronounced word with either of two visual targets. Infants showed significant preferences for correctly pronounced targets only. This suggests that the infants did not show a preference for vocalic or consonantal information in the word recognition task. However, our design does not allow us to determine whether the infants rejected both of the mispronounced words in each trial or whether they accepted both because they ignored both consonantal and vocalic information. We do consider rejection of both targets more likely, because same-age Danish-learning infants, as noted earlier, showed a bias toward better processing of vocalic information than consonantal information in a more demanding word-learning task (Højen and Nazzi 2016), and so they would seem unlikely to ignore this information in the less demanding word recognition task of the present study.
It should also be noted that the nonsignificant results do not permit a firm conclusion that there is no bias at all in Danish 20-month-olds’ familiar word recognition. Rather the results suggest, that should biases exist (see trends in Figure 3), their magnitude is likely negligible (η 2 < 0.01).
The overall picture of Danish 20-conmonth-olds’ use of consonantal and vocalic information in lexical processing (vowel bias in word learning but no bias in word recognition) is consistent with the hypothesis (Fennell and Werker 2003; Stager and Werker 1997) that infants selectively focus attention on certain aspects of the acoustic signal when processing demands are high (vowels in the case of Danish word learning, Højen and Nazzi 2016) but seemingly use more detailed information (both vowels and consonants) in a less demanding task (word recognition, this study) at the same age. This result is also consistent with a transition to consonant bias that is observed in native Danish adults’ processing of familiar words (Højen et al. in preparation). Adults have a fully developed and robust perceptual system and may focus attention on the most informative acoustic information, presumably consonant information, rather than the easiest information to process, presumably vowel information. The conflicting results regarding bias in processing of new words by native Cantonese adults may thus also result from differences in task difficulty (Gómez et al. 2018; Poltrock et al. 2018).
The pattern of results for Danish thus fits the general pattern for languages that have been examined in relation to lexical processing bias so far (see Figure 1 as well as Nazzi et al. [2016] and Nazzi and Cutler [2019]). Infants initially focus attention on vowel information but transition to consonant information at a point in development that depends on the characteristics of the language; perhaps acoustic-phonetic properties and/or distributional characteristics of the developing lexicon (for a discussion, see Nazzi and Cutler 2019). A specific hypothesis proposed by Chen et al. (2021) regarding the vowel bias found relatively late in development in Cantonese-learning children (30 months) is that Cantonese being a tone language, favors a focus on vocalic segments, which carry phonological tone variation. Including phonological tones multiplies the number of phonologically contrasting vowels and thus the number of potential lexical contrasts involving vowels. Moreover, vowels carrying tone variation may have an attentional effect favoring vowels. The suggestion by Chen et al. (2021) is consistent with the vowel bias found in native adults’ processing of familiar word in Mandarin, which is also a tone language (Wiener and Turnbull 2016; Wiener 2020), and partially consistent with the lack of a (consonant) bias found in native adults’ learning of new words in Cantonese (Poltrock et al. 2018) and the vowel bias in segmentation of new words (Gómez et al. 2018). The suggestion is also consistent with the relatively late transition from vowel bias to consonant bias seen in Danish, because Danish has both phonological vowel length and “stød” (which may also be viewed as a tone change) as noted in Section 1.1, which multiply the number of contrasts possible to make using vowel information and which may generally draw attention to vowel segments.
This explanation of vowel bias may tie in with the initial vowel bias in French and Italian infants during the first half-year of life (Nazzi et al. 2016), when infant-directed speech is characterized by exaggerated pitch excursions, a stretched vowel space and prolonged vowels (e.g., Kuhl et al. 1997), in other words, very vocalic speech. Infants are likely to hear utterances such as Yeeeees, spoooooon! (rather than Yesssss, spoonnnnn!). Moreover, until around seven months, typically developing infants show vocalic babbling without consonants (Eilers and Oller 1994). This highly vocalic environment in the first half-year in terms of both input and own production may induce the initial vowel bias in the first few months of life evidenced in French and Italian, and perhaps in all infants. When babbling begins to contain consonants after the first half-year and later becomes language-specific (de Boysson-Bardies et al. 2009), and when the speech infants hear becomes gradually more complex, the specific characteristics of the ambient language may favor different processing biases. Specifically, rather more consonantal languages like Italian and French may favor an early transition to a consonantal lexical processing bias, while for more vocalic languages like Danish, a consonantal bias may be deterred or delayed.
4.1 Limitations and future directions
An obvious limitation to the present study is the highly limited number of words examined. This is an unfortunate, but unavoidable, limitation because of the design. There are only so many words in the vocabularies of 20-month-olds that lend themselves to the type of paired reconstruction used here. Another limitation is the limited number of children tested. A lack of significant difference should be interpreted cautiously, as discussed. Although the effect sizes smaller than 0.01 indicate that a bias, should it exist, is very mild, a larger sample size would have strengthened the conclusions drawn from this study. These limitations can be addressed in future studies with different designs and larger sample sizes.
In addition, future research is needed to establish how biases change for processing of familiar words and new words in Danish. For example, it is unknown when the symmetry in bias seen in 20-month-olds’ processing of familiar words changes to a consonant bias that is observed in adults. In general, the multiple “blanks” in Figure 1 concerning lexical processing of familiar or new words at different ages and in different (and many more) languages need to be filled in by new research.
In summary, this study showed that Danish-learning 20-month-olds recognize familiar words with correct pronunciation but not when mispronounced. In addition, they showed neither a consonant bias nor a vowel bias in word recognition when presented with mispronounced familiar words that could be reconstructed by ignoring/changing a vowel or a consonant mispronunciation. The data for this research is available on OSF using this link: https://osf.io/u5sje/?view_only=006600caf1ec4d0fb5cd5052e2810914.
Acknowledgments
This research was supported by a grant from Widex A/S to the first author. We thank all participating families for their help in making this research possible, as well as two reviewers for helpful feedback and suggestions for improvement.
-
Author contributions: Anders Højen: Conceptualization, Methodology, Testing, Validation, Formal analysis, Investigation, Writing – Original Draft, Writing – Review & Editing, Visualization. Thomas O. Madsen: Conceptualization, Methodology, Writing – Original Draft, Writing – Review & Editing. Dorthe Bleses: Conceptualization, Methodology, Writing – Review & Editing.
-
Conflict of interest: The authors have no conflicts of interest to declare.
-
Ethics statement: Written informed consent was obtained from all families in the study. The study complies with the Helsinki Declaration. Non-invasive experimental research of the type conducted in the present study did not require approval from an ethics committee at the time when the research was conducted.
References
Acha, Joana & Manuel Perea. 2010. On the role of consonants and vowels in visual-word processing: Evidence with a letter search paradigm. Language and Cognitive Processes 25(3). 423–438. https://doi.org/10.1080/01690960903411666.Search in Google Scholar
Bailey, Todd M. & Kim Plunkett. 2002. Phonological specificity in early words. Cognitive Development 17(2). 1265–1282. https://doi.org/10.1016/s0885-2014(02)00116-8.Search in Google Scholar
Ballem, Kate D. & Kim Plunkett. 2005. Phonological specificity in children at 1;2. Journal of Child Language 32(1). 159–173. https://doi.org/10.1017/s0305000904006567.Search in Google Scholar
Benavides-Varela, Silvia, Jean-Rémy Hochmann, Francesco Macagno, Marina Nespor & Jacques Mehler. 2012. Newborn’s brain activity signals the origin of word memories. Proceedings of the National Academy of Sciences of the United States of America 109(44). 17908–17913. https://doi.org/10.1073/pnas.1205413109.Search in Google Scholar
Bleses, Dorthe, Werner Vach, Malene Slott, Sonja Wehberg, Pia Thomsen, Thomas O. Madsen & Hans Basbøll. 2008. The Danish communicative developmental inventories: Validity and main developmental trends. Journal of Child Language 35(3). 651–669. https://doi.org/10.1017/s0305000907008574.Search in Google Scholar
Bonatti, Luca L, Marcela Peña, Marina Nespor & Jacques Mehler. 2005. Linguistic constraints on statistical computations the role of consonants and vowels in continuous speech processing. Psychological Science 16(6). 451–459. https://doi.org/10.1111/j.0956-7976.2005.01556.x.Search in Google Scholar
Bouchon, Camillia, Caroline Floccia, Thibaut Fux, Martine Adda‐Decker & Thierry Nazzi. 2015. Call me Alix, not Elix: Vowels are more important than consonants in own‐name recognition at 5 months. Developmental Science 18(4). 587–598. https://doi.org/10.1111/desc.12242.Search in Google Scholar
Caramazza, Alfonso, Doriana Chialant, Rita Capasso & Gabriele Miceli. 2000. Separable processing of consonants and vowels. Nature 403(6768). 428–430. https://doi.org/10.1038/35000206.Search in Google Scholar
Carreiras, Manuel, Jon A. Dunabeitia & Nicola Molinaro. 2009. Consonants and vowels contribute differently to visual word recognition: ERPs of relative position priming. Cerebral Cortex 19(11). 2659–2670. https://doi.org/10.1093/cercor/bhp019.Search in Google Scholar
Carreiras, Manuel & Cathy J. Price. 2008. Brain activation for consonants and vowels. Cerebral Cortex 18(7). 1727–1735. https://doi.org/10.1093/cercor/bhm202.Search in Google Scholar
Chen, Hui, Daniel T. Lee, Zili Luo, Regine Y. Lai, Hintat Cheung & Thierry Nazzi. 2021. Variation in phonological bias: Bias for vowels, rather than consonants or tones in lexical processing by Cantonese-learning toddlers. Cognition 213. 104486. https://doi.org/10.1016/j.cognition.2020.104486.Search in Google Scholar
Creel, Sarah C., Richard N. Aslin & Michael K. Tanenhaus. 2006. Acquiring an artificial lexicon: Segment type and order information in early lexical entries. Journal of Memory and Language 54(1). 1–19. https://doi.org/10.1016/j.jml.2005.09.003.Search in Google Scholar
Cutler, Anne, Núria Sebastián-Gallés, Olga Soler-Vilageliu & Brit van Ooijen. 2000. Constraints of vowels and consonants on lexical selection: Cross-linguistic comparisons. Memory and Cognition 28(5). 746–755. https://doi.org/10.3758/bf03198409.Search in Google Scholar
de Boysson-Bardies, Bénédicte, Pierre Halle, Laurent Sagart & Catherine Durand. 2009. A crosslinguistic investigation of vowel formants in babbling. Journal of Child Language 16(1). 1–17. https://doi.org/10.1017/S0305000900013404.Search in Google Scholar
Delle Luche, Claire, Caroline Floccia, Lionel Granjon & Thierry Nazzi. 2017. Infants’ first words are not phonetically specified: Own name recognition in British English-learning 5-month-olds. Infancy 22(3). 362–388. https://doi.org/10.1111/infa.12151.Search in Google Scholar
Delle Luche, Claire, Silvana Poltrock, Jeremy Goslin, Boris New, Caroline Floccia & Thierry Nazzi. 2014. Lexical access in the auditory modality: Differential processing of consonants and vowels. Journal of Memory and Language 72(1). 1–5. https://doi.org/10.1016/j.jml.2013.12.001.Search in Google Scholar
Eilers, Rebecca E. & D. Kimbrough Oller. 1994. Infant vocalizations and the early diagnosis of severe hearing impairment. The Journal of Pediatrics 124(2). 199–203. https://doi.org/10.1016/S0022-3476(94)70303-5.Search in Google Scholar
Fennell, Christopher T. & Janet F. Werker. 2003. Early word learners’ ability to access phonetic detail in well-known words. Language and Speech 46(2/3). 245–264. https://doi.org/10.1177/00238309030460020901.Search in Google Scholar
Fenson, Larry, Virginia A. Marchman, Donna Thal, Philip S. Dale, J. Steven Reznick & Elizabeth Bates. 2007. MacArthur-bates communicative development inventories: User’s guide and technical manual, 2nd edn. Baltimore, MD: Brookes.10.1037/t11538-000Search in Google Scholar
Floccia, Caroline, Thierry Nazzi, Claire Delle Luche, Silvana Poltrock & Jeremy Goslin. 2014. English-learning one- to two-year-olds do not show a consonant bias in word learning. Journal of Child Language 41(5). 1085–1114. https://doi.org/10.1017/s0305000913000287.Search in Google Scholar
Gómez, David M., Peggy Mok, Mikhail Ordin, Jacques Mehler & Marina Nespor. 2018. Statistical speech segmentation in tone languages: The role of lexical tones. Language and Speech 61(1). 84–96. https://doi.org/10.1177/0023830917706529.Search in Google Scholar
Grønnum, Nina, Miguel Vazquez-Larruscaín & Hans Basbøll. 2013. Danish stød: Laryngealization or tone. Phonetica 70(1–2). 66–92. https://doi.org/10.1159/000354640.Search in Google Scholar
Havy, Mélanie & Thierry Nazzi. 2009. Better processing of consonantal over vocalic information in word learning at 16 months of age. Infancy 14(4). 439–456. https://doi.org/10.1080/15250000902996532.Search in Google Scholar
Havy, Mélanie, Josette Serres & Thierry Nazzi. 2014. A consonant/vowel asymmetry in word-form processing: Evidence in childhood and in adulthood. Language and Speech 57(2). 254–281. https://doi.org/10.1177/0023830913507693.Search in Google Scholar
Hochmann, Jean-Rémy, Silvia Benavides-Varela, Marina Nespor & Jacques Mehler. 2011. Consonants and vowels: Different roles in early language acquisition. Developmental Science 14(6). 1445–1458. https://doi.org/10.1111/j.1467-7687.2011.01089.x.Search in Google Scholar
Hochmann, Jean-Rémy, Silvia Benavides-Varela, Ana Fló, Marina Nespor & Jacques Mehler. 2018. Bias for vocalic over consonantal information in 6-month-olds. Infancy 23(1). 136–151. https://doi.org/10.1111/infa.12203.Search in Google Scholar
Højen, Anders & Thierry Nazzi. 2016. Vowel bias in Danish word‐learning: Processing biases are language‐specific. Developmental Science 19(1). 41–49. https://doi.org/10.1111/desc.12286.Search in Google Scholar
Højen, Anders, Signe Rømer Haubjerg, Boris New, Ocke-Schwen Bohn & Thierry Nazzi. in preparation. Consonant bias in Danish word reconstruction in adults.Search in Google Scholar
Kuang, Jianjing. 2017. Covariation between voice quality and pitch: Revisiting the case of Mandarin creaky voice. The Journal of the Acoustical Society of America 142(3). 1693–1706. https://doi.org/10.1121/1.5003649.Search in Google Scholar
Kuhl, Patricia K., Jean E. Andruski, Inna A. Chistovich, Ludmilla A. Chistovich, Elena V. Kozhevnikova, Viktoria L. Ryskina, Elvira I. Stolyarova, Ulla Sundberg & Francisco Lacerda. 1997. Cross-language analysis of phonetic units in language addressed to infants. Science 277(5326). 684–686. https://doi.org/10.1126/science.277.5326.684.Search in Google Scholar
Ladefoged, Peter & Ian Maddieson. 1990. Vowels of the world’s languages. Journal of Phonetics 18(2). 93–122. https://doi.org/10.1016/S0095-4470(19)30396-1.Search in Google Scholar
Ladefoged, Peter & Ian Maddieson. 1996. The sounds of the world’s languages. Oxford: Blackwell.Search in Google Scholar
Mani, Nivedita & Kim Plunkett. 2007. Phonological specificity of vowels and consonants in early lexical representations. Journal of Memory and Language 57. 252–272. https://doi.org/10.1016/j.jml.2007.03.005.Search in Google Scholar
Mani, Nivedita & Kim Plunkett. 2008. Fourteen-month-olds pay attention to vowels in novel words. Developmental Science 11(1). 53–59. https://doi.org/10.1111/j.1467-7687.2007.00645.x.Search in Google Scholar
Mani, Nivedita & Kim Plunkett. 2010. Twelve-month-olds know their cups from their keps and tups. Infancy 15(5). 445–470. https://doi.org/10.1111/j.1532-7078.2009.00027.x.Search in Google Scholar
Nazzi, Thierry. 2005. Use of phonetic specificity during the acquisition of new words: Differences between consonants and vowels. Cognition 98. 13–30. https://doi.org/10.1016/j.cognition.2004.10.005.Search in Google Scholar
Nazzi, Thierry & Anne Cutler. 2019. How consonants and vowels shape spoken-language recognition. Annual Review of Linguistics 5(1). 25–47. https://doi.org/10.1146/annurev-linguistics-011718-011919.Search in Google Scholar
Nazzi, Thierry, Caroline Floccia, Bérangère Moquet & Joseph Butler. 2009. Bias for consonantal information over vocalic information in 30-month-olds: Cross-linguistic evidence from French and English. Journal of Experimental Child Psychology 102(4). 522–537. https://doi.org/10.1016/j.jecp.2008.05.003.Search in Google Scholar
Nazzi, Thierry & Linda Polka. 2018. The consonant bias in word learning is not determined by position within the word: Evidence from vowel-initial words. Journal of Experimental Child Psychology 174. 103–111. https://doi.org/10.1016/j.jecp.2018.05.011.Search in Google Scholar
Nazzi, Thierry, Silvana Poltrock & Katie von Holzen. 2016. The developmental origins of the consonant bias in lexical processing. Current Directions in Psychological Science 25(4). 291–296. https://doi.org/10.1177/0963721416655786.Search in Google Scholar
Nespor, Marina, Marcela Peña & Jacques Mehler. 2003. On the different roles of vowels and consonants in speech processing and language acquisition. Lingue e Linguaggio 2. 221–247.Search in Google Scholar
New, Boris, Verónica Araùjo & Thierry Nazzi. 2008. Differential processing of consonants and vowels in lexical access through reading. Psychological Science 19(12). 1223–1227. https://doi.org/10.1111/j.1467-9280.2008.02228.x.Search in Google Scholar
Poltrock, Silvana, Hui Chen, Celia Kwok, Hintat Cheung & Thierry Nazzi. 2018. Adult learning of novel words in a non-native language: Consonants, vowels, and tones. Frontiers in Psychology 9. 1–16. https://doi.org/10.3389/fpsyg.2018.01211.Search in Google Scholar
Poltrock, Silvana & Thierry Nazzi. 2015. Consonant/vowel asymmetry in early word form recognition. Journal of Experimental Child Psychology 131. 135–148. https://doi.org/10.1016/j.jecp.2014.11.011.Search in Google Scholar
Ratnage, Paul, Thierry Nazzi & Caroline Floccia. 2023. Vowels and consonants matter equally to British English-learning 11-month-olds’ familiar word form recognition. Journal of Child Language 1–24. https://doi.org/10.1017/S0305000923000223.Search in Google Scholar
Stager, Christine L. & Janet F. Werker. 1997. Infants listen for more phonetic detail in speech perception than in word-learning tasks. Nature 388(6640). 381–382. https://doi.org/10.1038/41102.Search in Google Scholar
Swingley, Daniel & Richard N. Aslin. 2000. Spoken word recognition and lexical representation in very young children. Cognition 76(2). 147–166. https://doi.org/10.1016/s0010-0277(00)00081-0.Search in Google Scholar
Swingley, Daniel & Richard N. Aslin. 2002. Lexical neighborhoods and the word-form representations of 14-month-olds. Psychological Science 13(5). 480–484. https://doi.org/10.1111/1467-9280.00485.Search in Google Scholar
van Ooijen, Brit. 1996. Vowel mutability and lexical selection in English: Evidence from a word reconstruction task. Memory & Cognition 24(5). 573–583. https://doi.org/10.3758/bf03201084.Search in Google Scholar
Von Holzen, Katie & Thierry Nazzi. 2020. Emergence of a consonant bias during the first year of life: New evidence from own-name recognition. Infancy 25(3). 319–346. https://doi.org/10.1111/infa.12331.Search in Google Scholar
Wewalaarachchi, Thilanga D., Liang Hui Wong & Leher Singh. 2017. Vowels, consonants, and lexical tones: Sensitivity to phonological variation in monolingual Mandarin and bilingual English–Mandarin toddlers. Journal of Experimental Child Psychology 159. 16–33. https://doi.org/10.1016/j.jecp.2017.01.009.Search in Google Scholar
Wiener, Seth. 2020. Second language learners develop non-native lexical processing biases. Bilingualism: Language and Cognition 23(1). 119–130. https://doi.org/10.1017/S1366728918001165.Search in Google Scholar
Wiener, Seth & Rory Turnbull. 2016. Constraints of tones, vowels and consonants on lexical selection in Mandarin Chinese. Language and Speech 59(1). 59–82. https://doi.org/10.1177/0023830915578000.Search in Google Scholar
© 2023 Walter de Gruyter GmbH, Berlin/Boston
