Professional Documents
Culture Documents
Consonant Acquisition in Typically Developing Preschool Speakers - Raymond (2017)
Consonant Acquisition in Typically Developing Preschool Speakers - Raymond (2017)
OF GRANADA SPANISH
by
Gabriela Raymond
MASTER OF SCIENCE
in
(Vancouver)
April 2017
ABSTRACT
Spanish - a highly restricted coda dialect - were analyzed to determine the order of acquisition of
consonant singletons when influenced by word stress and word position of the consonants.
Words were elicited through a picture-naming task. Relative mastery of onsets in word-initial
and word-medial position were determined for left-, centre- and final-prominent words and sub-
categorized by their position in stressed and/or unstressed syllables. The present data concur in
general with existing literature on singleton acquisition in Spanish, but differ in terms of later
multisyllabic words. The most common general mismatch pattern was consonant substitution.
Both match and mismatch patterns are presented in detail by age group. The present study
supplements the Bernhardt’s et al. (2015) findings on word structure development in the same
cohort, and suggests further research for this and other dialects.
2
DEDICATION
To God for providing me with strength and hope in this long academic journey.
To my daughter Dani who has been my favourite associate and has sacrificed her life for me
To my love Blair who has been my greatest supporter and perpetual source of confidence.
To my dear friends who have been the best cheerleaders and have always helped me.
Very specially, to my loving parents and caring siblings because you created and shaped the
3
INTRODUCTION
Overview
Research on the acquisition of Spanish has focused mainly on the segmental level (Bosch, 1983;
Jiménez, 1987; Vivar & León, 2009; Henricksen & Willis, 2010; Susaníbar et al., 2013) with
minimal discussion of potential word structure (word length, word shape and stress as e.g
Bernhardt et al., 2015). The latter study analyzed single word samples of Spanish-speaking
preschoolers from Granada with typical (TD) and protracted phonological development (PPD1),
providing information on word length, shape and stress as well as segmental slots (timing units)
for singleton consonants and consonant sequences; the paper pointed out the need to create
norms for word structure acquisition and do further investigation on word structure-segment
interactions. The goal of the present study was to address this data gap: using the same data, to
determine the order of acquisition of consonant singletons when influenced by word stress and
Spanish dialectal variations are important to consider when planning assessment and/or
intervention. Different from many other varieties of Spanish, although similar to some Latin
American varieties, Granada Spanish is a language with optional codas. Given the coda-optional
quality of Granada Spanish, the focus for the current paper is onset consonants (word-initial, i.e.
WI, and word-medial intervocalic, i.e. WM). General information on Spanish phonology and
1
Protracted Phonological Development (PPD) refers mainly to the additional time a child needs to acquire the
phonology of a language when compared to those with typical development (Bernhardt & Stemberger, 1998).
Dubasik and Ingram (2013) referred to PPD as “typical phonological patterns persisting beyond expected ages,” and
they mentioned that speech intelligibility was lower than in those with typical development (<70%). PPD may be
referred to in the literature as speech sound disorder, speech delay, developmental phonological disorder or
impairment.
4
consonant acquisition is provided as background below, including characteristics of coda-
optional dialects.
Spanish Phonology
Spanish, the second most spoken language in the world (Ethnologue, 2017), is spoken both in
Europe and Latin America, with many dialectal variations. Dialectal studies have been done
about phoneme realization by context, i.e. their relation to the surrounding vowels and
consonants, the stress and shape of the syllable and the word, etc. (Moreno & Mariño, 1998;
Moya & García, 2010; Bernhardt et al., 2015; Oropeza, 2016). As noted by Alvarez (1987),
speaker education, socioeconomic status and geographical areas all contribute to these dialectal
differences that may appear in a language's semantics (e.g. different names for the same object),
phenomena that affect the production of specific phonemes). Some of these Spanish dialects and
particularly that of Granada, share some within-phoneme variation (Jiménez, 1987; Acevedo,
1993; Moreno & Mariño, 1998), i.e. allophonic variation. For example, there is variation
between voiced stops and fricative/approximants. Cataño, Barlow and Moyna (2009) classify [b,
Table 1 on the next page shows the consonant phonemes and allophones for Granada
Spanish.
5
Table 1. Consonant inventory of Granada Spanish
Plosives [-continuant] p ba t db k gc
Approximants [+sonorant] βa ðb ʝd ʝd ɣ c
Fricatives [-sonorant] f s/θ (ʃ)e xf
[+continuant]
Glides w j j w hf
[-consonantal]
a-g
Allophonic variants are labelled with matching letters.
Note. The allophone [h] can also substitute for /s/ in coda position. Glides /w/ and /j/ are the first elements of rising
diphthongs rather than consonants. The palatals have both [Coronal] and [Dorsal] place features. The parentheses
indicate less common variants.
Pertinent to allophonic variation for Granada Spanish, Chávez-Peón et al. (2012, p. 256) note the
following:
(c) the orthographic 'll' is [d͡ ʒ] word initially or after nasals, and [ʝ] or [j] in other contexts;
6
(g) [s] and [θ] present variability across speakers; some use ceceo, i.e. [θ], and some, seseo, i.e.
[s]; and
(h) nasals adopt the next consonant’s place of articulation by a process of assimilation when
Another variation concerns intervocalic /d/, often omitted in Granada Spanish when it is
the onset of an unstressed syllable that is preceded by a stressed syllable, and when it is part of
the suffix of a participle, e.g. caminado /ka.mi.ˈna.ðo/ > [ka.mi.ˈna.o] ‘walked’ (Moya & García,
2010). This phenomenon is observed in other Asturian communities and in some Latin American
Word structure
Word length in Spanish is varied. Disyllabic words are the most frequent, about twice as frequent
as multisyllabic and monosyllabic words (Bernhardt et al., 2015). Word stress can affect any
syllable. Disyllabic words can have right prominence (unstressed-Stressed or iambic, e.g. reloj
/re.ˈlox/ ‘clock’) but left prominence is more common (Stressed-unstressed or trochaic, e.g.
gorra /ˈgo.ra/ ‘cap’). Word stress is relevant to the present research because better consonant
accuracy is expected in stressed than unstressed syllables (Bernhardt & Stemberger, 1998;
of syllables (Justicia et al., 1996), with both diphthongs and consonant clusters. The syllable
combinations most commonly found in Spanish are: CV, CVC, V, VC, CVV and CCV (Justicia
et al., 1996). Consonant sequences may appear word initially and medially; they are not
commonly used in the same syllable (e.g. CCVCC) and may happen heterosyllabically with
7
sonorants and fricatives (e.g. /mb(ɾ)/, /n{θ/s}/) (Chávez-Peón et al., 2012). Four consonants are
allowed consecutively if they occur across syllable boundaries (heterosyllabic, e.g. transcribir
‘to transcribe’ CCVCC.CCV.CVC). Spanish words may end in any of the five Spanish vowels
/a, e, i, o, u/ or the following single codas:{d/ð}, {n /ŋ}, /s, x, l, ɾ/. Word-final codas present
Acquisition of phonemes is a process that starts at a very early age when infants listen to the
language and start exploring its use with vocalizations. Bosch (1983) synthesized different
process that happens over long periods of time with 'correct' and 'incorrect' production of those
targets; and (b) some groups of phonemes are acquired earlier than others, e.g. early-developing
voiceless stops and nasals versus later-developing fricatives, liquids and trills (Melgar de
González, 1976; Jiménez, 1987; Vivar & León, 2009). However, knowledge of this trend was
not assumed to determine all children’s developmental path or imply that some speech sounds
are necessary foundations for developing others. More precisely, the speech sounds and
structures are learned in different contexts in relation to each other. General ideas about
acquisition only allow us to compare children's productions with norms but do not in fact
describe a child's actual phonological development (Bosch, 1983). When children are not able to
produce some sounds or structures, they may simplify words by changing the syllabic structure,
8
A number of researchers have examined consonant development for different Spanish
dialects. Some major ones chronologically include Bosch for Castilian Spanish (1983, 2004);
Jiménez for California Spanish (1987); González for Andalusian Spanish (Málaga, 1989);
Acevedo for California Spanish (1993); and Vivar and León for Chilean Spanish (2009).
First, for Castilian Spanish, Bosch (1983) designed a test using 32 high frequency words
including monosyllables, disyllables with initial (trochaic, left-prominent) stress or final (iambic,
right-prominent) stress, and trisyllabic words (3 with word-medial, centre-prominent stress and
one with word-final stress). The words were obtained by spontaneous elicitation or if necessary,
by imitation. Participants included 293 children between the ages of 3 and 7;11. Table 2 shows
children's percent correct consonant match by age. Many phonemes were acquired by age 3 but
9
Table 2. Percent consonant match (PCM)* by age in Castilian Spanish-speaking children
(Bosch, 1983, 2004) - adapted
90% m, n, ɲ d, g, f s, θ r
p, t, k, b ͡ tʃ
x, l ɾ
Early-acquired
consonants 80% g s s, θ ʎ ʎ
f, s ʎ ʎ r
͡ tʃ
ɾ
Intermediate 70% d θ r
stage r
consonants
60% ʎ
Late-acquired <50% θ
consonants r
* Table in original article refers to these results qualifying them as “onset or coda” but details of the phonemes in
each position were not provided.
During a previous investigation, Bosch (1983) observed phonological processes that appeared
frequently in the data at different ages; these were substitutions, assimilations and changes in
syllabic structure. She further categorized them as “normal,” “indicative of risk,” and “indicative
of severity” (Bosch, 1983, p. 106-109, paraphrased and translated). In Bosch (2004), she deleted
the third category from her age profile tables concerning phonological evaluation of child
speech. Here we mention those involving consonant singletons from her 2004 investigation. At
age 3 trill was not acquired, and common patterns were stridency of /θ/ (produced as [s, f]),
fronting of /s/, lateralization of intervocalic /ð/, and gliding of /ʎ/ (i.e. [j]) (Bosch, 1983, p. 106,
paraphrased and translated). The same was true for the 4-year-olds except for lateralization of
/ð/. At age 5, trill was still not acquired, and there was gliding of /ʎ/ (produced as [j]), and
stridency of /θ/ (produced as [s]). Regarding assimilations, the 3-year-olds were showed velar,
labial and nasal assimilations; the 4-year-olds, only velar. Assimilations were indicative of risk
10
for the older groups. Lastly for syllabic structure, reduction of clusters were common at age 3,
but only of /ɾ/-clusters at ages 4 and 5 years. Deletion of /ɾ/ and fricatives occurred at age 3 - the
latter being typical in Granada Spanish at all ages. At age 4, omission of coda /ɾ/, and mistakes
California between the ages of 3;0 and 5;7. The test involved elicitation of spontaneous single
words; delayed imitation was used in the absence of an answer. The test included 38 words
multisyllabic words (14 trisyllabic and 2 longer words) with different stress patterns. Dialectal
variations were scored as correct. The results of this study were presented in intervals of 4
months (e.g. 3;0-3;3). For the purpose of this study, the final stage for each age group proposed
by Jiménez will be presented as cumulative for the whole year period instead of the four months
(3;11, 4;11 and 5;7). (See Table 3.) Percentages were provided per sound by word position and
11
Table 3. Acquired consonants by age group (Jiménez, 1987) - adapted
p, t, k, b d g
m, n ɲ ͡ tʃ
Early-acquired
80-100% w f, x, s, j
consonants
l
ɾ
f g s
Intermediate stage ͡ tʃ ͡ tʃ
70-79% r
consonants
j
d, g r
ɲ
Late-acquired
consonants
<69% s, x
ɾ, r
l
Jiménez (1987) presents this data in a horizontal bar chart reporting a range from 50-90% consonant
accuracy. Information was extracted and adapted to fit this table.
Acevedo (1993) replicated Jiménez’ study for the same population. She obtained the same
findings for /m, n, w, p, b, t, k, ɾ, r, ͡tʃ/, but found earlier acquisition of /ɲ, d, f, s, x, j, l/ and later
acquisition of /g/. The author mentioned that the differences could have been influenced by some
methodology factors (e.g. single examiner in the majority of the children assessed instead of two
examiners like in the original study; or, the high number of imitated responses possibly due to
Turning to Andalusian Spanish (the context for the current study), González (1989)
studied 198 children aged 3 to 6;11 from the province of Málaga. She elaborated a screening test
that used an imitation task to register the children’s productions and the phonological processes
they used. One of her goals was to observe the Andalusian dialectal influence. Specific
phonemes acquired per age group are difficult to determine from the data presented. However,
she concludes that the earliest acquired singletons are “plosives, affricate, nasals and the
fricatives /f, x/, while the latest acquired are /s, θ/, liquids” (González, 1989, p. 22). She also
12
pointed out that children simplified their productions with: (a) substitutions (for plosives and /s,
θ, f, ͡tʃ, l, r/), occurring frequently across ages, followed by (b) syllable changes involving /s, θ/,
nasals, /l/ and /r/, and lastly, (c) assimilations for /ɾ, r/. At age 3, assimilations of singletons and
phonological processes affecting the syllable structure were uncommon. However, substitutions
of plosives, nasals and /t͡ ʃ , l/ appeared often. At age 6, the target 'll' was not fully developed,
hence substituted. González (1989) noted that in her population, phonological processes
affecting /s, θ, ʎ, r/ presented more persistently over time, possibly due to the influence played
aspiration, lateralization of /r/, as well as omissions of word-final phonemes such as /l, r, s/. The
current study on Granada Spanish will also discuss these characteristics in the Discussion.
Finally, relative to South American Spanish dialects, in 2009, Vivar and León studied 72
Chilean children aged 3 to 5;11 using a Cuestionario para la Evaluación Infantil (CEFI) with
high frequency words for children speaking the Chilean dialect. Phonemes were assessed in WI,
WM and WF position but results were condensed by phoneme. No analysis was done regarding
word position or word length. The tokens, however, included eight trisyllabic words and two
longer words. The Chilean dialect presents some dialectal variations: (a) omission of WF /d/ (e.g.
ciudad ‘city’ /siu.ˈda/), and (b) aspiration of /s/ in coda position word internally and word finally
(e.g. escoba ‘broom’ /eh.ˈko.ba/) (Vivar & León, 2007), both patterns also found in Granada
Spanish. The results were described in terms of six groups consisting of 6-month periods. For the
purpose of the present study, the results presented in table 4 consist of the second (later) period
per age, i.e. 3;6-3;11, 4;6-4;11, and 5;6-5;11. Vivar and León (2009) conclude that their results
2
“Seseo” refers to the production of the fricative /s/ for orthographic 'c', 's', and 'z' in all contexts.
“Ceceo” refers to the production of /θ/ for orthographic 'c' before the letters 'e' and 'i' at least and
sometimes more pervasively.
13
concur with previous research in acquisition of phonemes: nasals and /b, d, ʝ, g/ showed a higher
percentage of correct production while /f, s, x, l, ɾ/ and especially /r/ showed more articulatory
difficulty, therefore later acquisition. The phonological processes that were most commonly
present were substitution, followed by modification and then omission. (The authors define
modification as the phonological process in which a feature of the phoneme is changed, e.g. /s/ >
/θ/. They do not allude to surrounding phonemes when this process is explained. Modifications
as referred by Vivar and León [2009] would be considered a type of substitution in the present
Table 4. Percent consonant match (PCM) * by age in children speaking the Chilean dialect
(Vivar and León, 2009) - adapted
100% p, t, b p, t, k, b p, t, k, b, d, g
m, n, ɲ m, n, ɲ m, n, ɲ
l l
ʝ r
ʝ
90% ʝ d, g s, f
Early-acquired ͡ tʃ
l f
consonants ͡ tʃ
ɾ
80% d, g s, x x
͡ tʃ r
f, x
ɾ
70% r r
Intermediate stage
consonants 60% s
Late-acquired <50%
consonants
14
Mason, Smith, and Hinshaw (1976) (cited in Jiménez, 1987) found that /n, h, f, d, s, ɾ, r/ were
mastered later than age 5 in children of Mexican lineage. Linares (1981, cited in Jiménez, 1987)
observed that /b, s, r/ were acquired after age 5 in Spanish monolinguals from Chihuahua,
Mexico, while /b, s, ɾ, r, ͡tʃ/ was pronounced correctly after age 5 in Spanish speakers (mono- and
bilingual) from New Mexico. Although not the focus of this study, it would be interesting to
The five main investigations described above (Bosch, 1983, 2004; Jiménez, 1987;
González, 1989; Acevedo, 1993; Vivar & León, 2009) confirm that acquisition of phonemes is a
gradual process. Participants in each study spoke different dialects of Spanish and the results had
many similarities. In general, nasals /m, n/ appeared earlier in acquisition with some variability
of /ɲ/ among studies, appearing even as late as age 5. Voiceless plosives were acquired most
often at age 3, with one study showing acquisition as late as age 4. The voiced plosive /b/ was
acquired at 3 but /d, g/ presented some variability, being acquired as late as age 5. The affricate
/t͡ ʃ/ was generally at an intermediate stage for age 3, mastered commonly at 4 years but
sometimes as late as 5 years. The fricatives /f, x/ were acquired earlier than /s, θ/ and the
pronunciation of letter tuplet 'll' (sometimes referred to as ʎ/) with the latter acquired even as late
as age 7. The liquids /l, ɾ/ were acquired earlier than /r/, the latter often acquired as late as age 7.
Studies agree that age 4 is a major milestone in acquisition. Table 5 below integrates data from
15
Table 5. Summary of relative mastery for the five studies described
(Bosch, 1983, 2004; Jiménez, 1987; González, 1989; Acevedo, 1993; Vivar & León, 2009)
m, n ɲ d, g s, θ ʎ
p, t, b k ͡ tʃ r
f, x
l, ɾ
The question that still remains is about the role of word structure in acquisition. It appears that
researchers considered this factor because they elicited the speech sounds in different word
positions, lengths and stress patterns but no analysis is offered relative to these variables.
Jiménez and Acevedo (1993) provided some information but did not discuss how word position
impacted mastery of the speech sounds. Other research in different languages (English -
Grunwell, 1985; Peruvian Spanish - Susaníbar, Huamaní & Dioses, 2013; Portuguese - Yavaʂ,
Hernandorena & Lamprecht, 2001, cited in Susaníbar et al., 2013) noted that the position of the
singletons in the word influences acquisition, as does the structure of the syllable (Bosch, 1983;
Bosch, 2004; Bernhardt et al., 2015). Later on, Bernhardt et al. (2015) expanded on word length
and stress, all factors that contribute to the correct production of the sounds in the word.
The frequency with which phonemes appear in a specific language or dialect is another
factor that affects acquisition (Bosch, 2004; Vivar & León, 2009). González (1989) set out to
demonstrate how age, gender and socio-economic status affect phoneme acquisition, pointing out
that many phonological differences presented by children from Málaga (in Andalusia, just like
Granada) were typical of the Andalusian dialect and not errors of production. She explained that
despite the dialect, the phonological development followed the same trajectory as Castilian
(North Central) European Spanish, but was “masked due to the dialectal difference” (González,
16
1989. p.22). She suggested further investigation in order to analyze typical errors produced by
children speaking the Andalusian dialect. This present research intends to describe that natural
As we study the acquisition of sounds in different languages, it is easy to focus only on the
segments and their distinctive features. The position of segments in words and the prominence of
the syllable in which the phonemes exist are some factors that determine the behaviour of that
segment. For example, a child may be able to produce a phoneme in onset position but not in
coda. The surrounding phonemes also affect the ability of the child to produce a specific sound.
The length of a word from mono- to multisyllabic represents more difficulty as the number of
syllables increases. It could manifest in ways in which the child inserts or deletes timing units. In
general, if a phoneme belongs to an unstressed syllable, its production becomes complex; if that
unstressed syllable is located word initially, then the complexity is higher (Bernhardt &
Stemberger, 2000). Word shapes can either facilitate or complicate the production of some
phonemes. For example, the second segment of a CCV shape tends to show more phonological
Children produce words in certain ways depending on the interaction of all these
elements involving features, segments, position of segment and the sequences in which it exists,
and word structures elements such as length, stress and word shape. All these factors must be
considered when trying to determine the acquisition of phonemes. Bernhardt et al. (2015)
analyzed how word length and stress resulted in accurate segment productions or in mismatches
in Granada Spanish. As with other languages, they observed that phonemes that appeared in
17
unstressed syllables were likely to be deleted, and more so if the phoneme was in the first
syllable of words that had centre prominence. The relationship between word length and
structure was observed when in disyllabic syllables of unstressed-Stressed shape (uS), the first
syllable did not get deleted but it did when the word was longer unstressed-Stressed-unstressed
(uSu). So the ability to pronounce a phoneme requires more accuracy when it falls in an
1. Consonant accuracy was expected to increase from age 3 to 5 years, i.e. the older the children,
the higher the accuracy (following Bosch, 1983, 2004; Jiménez, 1987; González, 1989;
2. Better consonant accuracy was expected in stressed than unstressed syllables (Bernhardt &
Stemberger, 1998; Chávez-Peón et al., 2012; Bernhardt et al., 2015) independent of word-
3. Trends for WI and WM position were expected to match those of research in other dialect
areas for:
(i) For matches with adult targets: acquisition of speech sounds of Granada Spanish would be
similar to table 5, summarizing mastery across dialects and especially Málaga Spanish; and
(ii) For mismatch patterns: substitutions would be the most common mismatch pattern,
followed by word structure effects, assimilation being the least common) (Bosch, 1983,
18
METHOD
Participants
The original study of Granada Spanish as reported in Bernhardt et al. (2015) included 59
participants divided into two groups: (a) a TD group of 30; and (b) a group with PPD of 29.
Their parents signed the consent forms received from the classroom teachers in accordance with
ethical agreements from the two universities involved. The selection criteria for the TD group
included: (a) normal hearing (hearing screening at 25 dB from 250 to 4000 Hz); (b) a normal oral
mechanism; and (c) language comprehension and production scores in the normal range on:
Prueba de lenguaje oral Navarra—Revisada (PLON-R3; Aguinaga et al., 2004) for 3-year-olds,
the Test de comprensión de estructuras gramaticales de 2 a 4 años (Calet et al., 2010); the
Peabody Picture Vocabulary Test—Español (Dunn et al., 2006); and the Test breve de
inteligencia de Kaufman (Kaufman & Kaufman, 2009; Spanish adaptation, Cordero & Cologne,
2000) (Bernhardt et al., 2015, p. 301). The PLON-R phonology subsection and a short dialogue
with the participant guided the assignment to the TD/PPD groups. After the initial group
assignment as TD/PPD, whole word match (WWM) was used as a secondary global measure to
finalize group assignment. This measure evaluated the children’s phonological productions in
response to a 103-word naming task (Bernhardt et al., 2015) study (see below under
Procedures). Whole word match was determined by the number of responses provided by the
child that matched the adult productions. Correct scoring was given to the variations
characteristic of the Granada dialect, and “slight phonetic deviations concerning voicing quality,
3
The PLON-R subsection includes mostly disyllabic trochees divided into 21 words for age 3, 23 for age
4, and 12 words for age 5. The emphasis of this phonology subsection is on acquisition of segments by
age group. The patterns obtained from each participant’s dialogue were compared based on age to Bosch
(2004) regarding typical and atypical patterns.
19
dentalization or vowel quality were ignored” (Bernhardt et al., 2015, p. 302). Means and
standard deviations based on age and group were obtained in order to compare the information
gathered from each child (see Table 6). This procedure followed research in English for WWM
(Schmitt et al.,1983). The final TD group included “ten 3-year-olds, nine 4-year-olds and eleven
5-year-olds (16, 14 boys: a higher proportion of girls at ages 3 and 4: 13/19)” (Bernhardt et al.,
2015, p. 302). Group assignment was confirmed per age according to Mann-Whitney U tests,
tests.
Table 6. Percent whole word match (WWM) and word shape match (WSM) for
Granada Spanish-speaking preschoolers with typical development
(adapted from Bernhardt et al., 2015)
20
Procedures
The Bernhardt et al. (2015) study used a list of 103 single words for a picture-naming task. (An
The present study included 86 words from the beginning list (see Appendix 2) consisting of 8
monosyllabic words, 54 bisyllabic, 23 trisyllabic, 10 tetrasyllabic and 1 word with five syllables.
(See Appendix 2 - List of words used). The data were collected by a native Spanish speaker and
audio-recorded with a Microtrack II digital recorder and associated wireless microphone. When a
spontaneous word was not obtained during the assessment, it was elicited through indirect
imitation: e.g. for hipopótamo 'hippopotamus' a question would be asked, such as “Is this a
hipopótamo or a puerta ('door')?”. If an answer was not obtained in this way, then the child was
The Granada and North America teams involved in this study developed narrow
transcription norms specific for the obtained data, where pronunciation differences produced by
adults were considered for each word. These procedures were developed for a crosslinguistic
investigation (Bernhardt & Stemberger, 2012). The first transcriptions were done for 12 children
in the PPD group not used for this study. It is relevant to mention it here because it was followed
by adjustments of the initial transcription norms. The TD data sets were transcribed by Granada
Spanish native speakers, and were confirmed after a second transcription done by the North
American team. For tokens in which there was lack of agreement, waveform and spectrogram
analysis was used. Two factors were analyzed relative to transcription agreement: timing unit
presence, and segments; the segments had to appear in the correct position, but if omitted, then
timing of the syllable was maintained by lengthening other segments. Agreement for the TD data
21
was 95.6% for timing unit presence, and 96% for segments. During a final meeting, consensus
Analysis
Rose and Hedlund’s (2013) Phon 1.5 software, and spreadsheets were used to analyze the
transcriptions qualitative and quantitatively. A match (correct) was the one that coincided with
adult productions, taking into consideration dialectal variation. For example, if a child did not
use an optional coda, the adult variant without a coda was assumed to be the child’s target. For
the specific purpose of this study, singleton consonant data were organized in spreadsheets as
follows: (a) by age group (3, 4 or 5 years); (b) by position in the word (WI, WM intervocalic,
and WF); and (c) by stress prominence (left, centre, right). Left prominence refers to
monosyllabic words, or those that start with a stressed syllable (e.g. luz /ˈlus/ 'light', gato /ˈga.to/
'cat', lámpara /ˈlam.pa.ɾa/ 'lamp'). Centre prominence indicates a word with the stressed syllable
in the middle (e.g. chimenea /t͡ ʃɪ.me.ˈne.a/ 'chimney', teléfono /te.ˈle.fo.no/ 'phone'). Right
prominence refers to words that end with a stressed syllable (e.g. ratón /ra.ˈton/ 'mouse'). No
statistical value was assigned due to the small dataset, but trends by percent match correct were
calculated. The formula used was [(# of words produced for a given consonant in a particular age
group and position / # of consonant matches of the same kind) x 100] = % consonant match).
The number of tokens per phoneme in different position and prominence varied depending on the
word list and what was obtained from the children (See Appendix 3 for number of tokens per
phoneme). Consonant mismatches were tallied in the same spreadsheets and categorized by the
types of errors produced by the children (e.g. substitution, deletion, epenthesis, etc.).
22
RESULTS
Results are presented by age group and word position (WI, WM, WF), and within this, by stress
prominence (left, centre, right). Age-specific data are presented in Tables 7, 8 and 9, one for each
of the age groups. Additional information presented in these age-specific tables includes: (a)
nonlinear set of features, as shown in Appendix 1, or the Nonlinear Scan Analysis form for
in voiced 'stops'/approximants grouped at the bottom of these tables (e.g. /β/ for globos
'balloons', /ð/ for ruido 'noise', /ɣ/ for agua 'water'); and (c) percentages of consonant match
where <50% indicates “emerging phoneme”; 51-70% indicates a low match categorized as
“phoneme still developing”; 70 - 85% indicates “phoneme near mastery level”; and 86 - 100% is
considered a “mastered phoneme.” Only two phonemes (i.e. /d/, /x/) were “absent” indicating
0% matching but are not considered as a different category because fewer than five tokens with
these phonemes were obtained. Given the focus of this study on consonant onsets, compiled
results are presented in tables 10 and 11 for WI and WM respectively. Finally, general mismatch
patterns are presented by age group in tables 12, 13 and 14. The asterisk (*) with which some
phonemes are marked indicates that five or fewer tokens were found in the sample. These results
should be considered with caution because they might not be fully representative of the sample.
23
Consonant Match by Age Group, Word Position and Stress (Prominence)
The summary of consonant acquisition is organized by age group, mastery level and sound
classes. In each of the mastery and non-mastery categories, sound class acquisition is described
first across onset positions, and then by WI, WM and WF only. Within each manner category,
notation about place of articulation and laryngeal features is noted. At the end of each age group
3-year-olds
Table 7 follows on the next page, prior to the discussion of the 3-year-old patterns so as
not to interrupt the flow of the detailed description of the various sub-sections.
24
25
3-year-olds: Left prominence
(*/d͡ ʒ/ in WM)
[+voiced], coronal [+anterior] [ð], and dorsal [ɣ]; and ([-consonantal]) [-voiced] [+spread
glottis] [h]
26
Near-mastery: left-prominent, WI stressed only
[-anterior][+grooved] /t͡ ʃ/
27
Still developing: left-prominent, WM unstressed only
28
3-year-olds -- Centre-prominent word-initial - unstressed syllables only
[-anterior][+grooved] /t͡ ʃ/
29
Near-mastery: centre-prominent, WM in both un/stressed syllables
30
2. Nasal ([+nasal]): [coronal [+anterior] /n/ and coronal-dorsal /ɲ/
31
3-year-olds -- Right-prominent: WM stressed syllables
One purpose of the study was to evaluate the possible effect of stress on consonant development.
The above analysis and Table 7 list the segments in detail by level of mastery, word position and
prominence. Overall, consonant mastery was relatively high and many segments appeared across
word positions and prominences. In order to compare with the older children, the following
possible or unlikely factor here and we return to this further in the Discussion.
32
First, trilled /r/ was still unmastered across all contexts (stress thus being irrelevant).
coronal and dorsal stops /d/ and /g/, the /g/ also in left-prominent stressed syllables; thus,
prominence was possibly irrelevant for /g/ but more so for /d/. Word medially, tap was still
developing in centre-prominent stressed syllables but was mastered or nearly mastered in all
other contexts, suggesting that prominence might have been at least in part relevant for tap.
Regarding fricatives, the /s/ was still developing word medially in stressed centre-prominent
contexts, also suggesting a possible effect of prominence. The allophonic approximants /ð/ and
/ɣ/ were unmastered word medially, the former in both left- and centre-prominent contexts, and
the latter in stressed centre-prominent contexts (stress perhaps less relevant than feature
development). Finally, the geminate /nn/ was unmastered in left-prominent stress contexts.
4-year-olds (Table 8)
Again, the tabular data for the 4-year-olds are presented ahead of the text, due to the detail in the
33
34
4-year-olds (Table 8)
1. Voiceless stops ([-continuant][-nasal]): [-voiced], all places /p, t, k/. (WM [+voiced]: /b, d, g/
3. Fricatives/approximants ([+continuant]):
35
4-year-olds -- Left-prominent: Word-initial stressed syllables only
/ʃ/*
[-anterior][+grooved] /t͡ ʃ/
[+anterior][-grooved] /θ/
36
4-year-olds: Centre prominence
syllables
allophones: [+voiced], labial [β] (*[β] for WI); coronal-dorsal [ʝ] (ʝ* for WI); and dorsal [ɣ]
allophone: [+voiced] dorsal [ɣ]; and ([-consonantal]): [-voiced] [+spread glottis] [h]
37
Centre-prominent word-medial un/stressed syllables
[+anterior][+grooved] /s/
[+anterior][+grooved] /s/
38
4-year-olds: Right prominence
1. Nasal ([+nasal]): coronal [+anterior] /n/ (WI /m/ was not elicited)
allophone ([-consonantal]): [-voiced] [+spread glottis] *[h] (WM /h/ was not elicited)
39
3. Voiceless fricatives ([+continuant][-sonorant]): [-voiced][+spread glottis], coronal
At age 4, most WI consonants were mastered. Exceptions were: (a) /b, g/, which were near
mastery in the unstressed context; (b) /r/. near-mastery in the left- and right-prominent words in
stressed and unstressed syllables respectively (e.g. reloj, rojo). Thus, for the voiced stops /b, g/,
stress appeared to have an expected influence, whereas this was not probably the case for the
trill.
Word medially, there was more inconsistency. Mastered categories included: (a)
stops and nasals, except for /m/ in unstressed syllables of centre-prominent words; (b) sonorants
tap /ɾ/, lateral /l/ and allophones [β], the latter in both unstressed and stressed syllables of left-
and centre-prominent words respectively, and also [ʝ] but in unstressed syllables ; (c). trill in
unstressed syllables of centre-prominent words (e.g. guitarra); (d) fricatives /s/ in both
unstressed and stressed syllables of left- and centre-prominent words respectively, /x/ only in
unstressed syllables of left- and centre-prominent words, and /θ/ in stressed syllables in words
40
Near mastery word medially were: (a) trill in unstressed syllables in words with left
prominence (e.g. perro); (b) fricative /θ/ in unstressed syllables of left-prominent words; and (c)
[ɣ] across prominence contexts. Stress appeared to play a somewhat relevant role in that early-
acquired phonemes such as /m/ were not mastered in the WI unstressed context, yet stress
probably was secondary or had no effects in the later-acquired such as [ɣ] or perhaps fricatives
and /r/.
5-year-olds - (Table 9)
The tabular data for the 5-year-olds are presented ahead of the text, due to the detail in the text.
41
42
5-year-olds: Left prominence
1. Stops ([-continuant][-nasal]): [-voiced], labial, coronal and dorsal /p, t, k/; (WM [+voiced] /b,
4. Approximant allophones [+voiced], labial [β]; coronal-dorsal [ ʝ]; dorsal [ɣ] (*[β, ʝ, ɣ] for WI)
43
5-year-olds --Left-prominent WM unstressed syllables only
[-grooved] /θ/; allophone [+voiced], coronal [+anterior] [ð]; and ([-consonantal]): [-voiced]
[-anterior][+grooved] /t͡ ʃ/
syllables
44
Near-mastery: centre-prominent both WI and WM un/stressed
1. Voiced stops ([-continuant][-nasal]): [+voiced], labial and coronal [+anterior] /b, d/ (not
elicited WM)
[+anterior][+grooved] /s/
45
3. Tap ([+vibrant][-trill]), coronal [+anterior] /ɾ/
dorsal [ð, ʝ]
46
5-year-olds -- Right-prominent: Word-initial unstressed syllables only
1. Nasals ([+nasal]): coronal [+anterior] /n/ (/m/ was not elicited in WI)
allophone ([-consonantal]): [-voiced][+spread glottis]: *[h] (/h/ was not elicited in WM)
47
Summary note on unmastered consonants by prominence in 5-year-olds
Results for age 5 in WI and WM position were very similar to those for age 4. The changes
observed word-initially were: (a) for [-continuant] voiced stops: /b/ went from nearly mastered to
mastered in unstressed syllables of centre-prominent words, but /g/ stayed in near-mastery; (b)
the trill /r/ went from nearly mastered to mastered in stressed syllables of left-prominent words,
but still appeared as nearly mastered in unstressed syllables of right-prominent words; and (c) the
of centre-prominent words but 5 or less than 5 tokens were elicited in the sample.
All WI phonemes showed mastery in stressed syllables of left-prominent words. The /g/
and /r/ were nearly mastered in unstressed syllables of centre-prominence and right-prominent
words respectively.
WM changes at this age were: (a) /θ/, [ɣ] and /r/ went from nearly mastered to mastered
in unstressed syllables of left-prominent words; (b) /m/, /s/ went from nearly mastered in
unstressed syllables of centre-prominent words, but [ɣ] regressed from mastered to nearly
mastered. In stressed syllables, however, [ɣ] moved from nearly mastered to mastered in both
centre- and right-prominent words. Once again, these results suggested that stress of the syllables
in which a phoneme appears is influential in the accuracy of its production. We return to this
The results presented above for ages 3, 4 and 5 have been condensed in tables 10 and 11.
They show the relative mastery for word-initial consonants by word position and prominence in
relation to age (Table 10) and for word-medial consonants (Table 11).
48
49
50
Consonant Mismatches or Dialectal Variants
The following sections describe the relative frequency of mismatch types for consonants and
features at the three ages (3, 4, 5 years). Within each age group, the data are organized by word
position (WI, WM) and prominence, (left, centre, right). The following abbreviations pertain:
Sub = substitution; Del = deletion; Lg = lengthen; Epen = epenthesis; Short = shortening; [ct] =
[continuant]; [cs] = [consonantal]; [lat] = [lateral]; [nas] = [nasal]; [vibtr] = [trill]; [vibtp] = [tap,
single vibrant]; [Lab] = [Labial]; [rd] = [round]; [Cor] = [Coronal]; [Dor] = [Dorsal]; [ant] =
[anterior]; [gd] = [grooved]; [vd] = [voiced]. Feature mismatches in the substitutions are
presented with the target feature value, and the resulting value, e.g. [-ant] → [+] (the > symbol
means that one substitution is more frequent than another). The allophonic variants of voiced
stops are considered approximants, i.e. [+sonorant]. Note that some of the substitutions may be a
Overall, consonant mismatches and acceptable dialectal variants were most frequent in
the younger group. The most common pattern was consonant substitution generally followed by
syllables occurred most often with multisyllabic words with centre prominence, with occasional
51
Mismatches and dialectal variants: Age 3 (Table 12)
52
Mismatches and dialectal variants: Age 3 (Table 12)
3-year-olds: WI position
The details below provide information on each consonant showing mismatches. As an overview,
although there were far fewer targets in centre-prominent words, there were more mismatch
frequently than substitutions). In fact, consonant substitution was the only mismatch pattern for
patterns occurred across prominence categories. Features most often showing changes were
consonants such as trill, tap, /s/, affricates, dorsals. Sometimes the substitutions only affected one
feature of a consonant but often there were multiple changes, e.g. [+vibrant] might appear as
[-vibrant] only (a tap) but also a [z], losing [+vibrant] but also [+sonorant], a [d], losing
[+sonorant][-continuant] or even a glottal stop, losing [C-Place] and [+sonorant]. The acceptable
variation between /s/ and /θ/ and rhotics and lateral in coda are included here as are allophonic
variation between voiced stops and approximants. They are indicated with AV for acceptable
variant.
/s/: C-Sub /s/ → [ɕ], [s̞ ], [ʃ] [+gd] → [-]; [+ant] → [-]
/g/: C-Sub /g/ → [b] > [gw], [l], [Dor] → [Lab/Lab]; > [-lat] → [+];
53
/r/: C-Sub /r/ > [d], [z], [ʔ] [+vibtr] [-lat] → [-vib] [-lat]; [+son] →
/t/: C-Sub > Syl-Del /t/ → [ts], [k] [-ct] → [-,+]; [Cor] → [Dor]
/s/: C-Sub > Syl-Del /s/ → [t] > [ʃ] [+ct] → [-] > [+ant] → [-];
/t͡ ʃ/: C-Sub > C-Del /t͡ ʃ/ → [ʃ] (AV) > [t] [-ct.+ct] → [+ct] (AV) > [-ct,+ct] → [-ct]
/x/: C-Sub > C-Del /x/ → [t], [f] [+ct][Dor] → [-ct][Cor]; [Dor] → [Lab];
/r/: C-Sub > C-Del /r/ → [l] > [g] > [ɹ] [-lat][+vibtr] → [+lat][-vib] >
[+cs][+vibtr] → [-cs][-vib]
54
[l] [-lat][Dor] → [+lat][Cor];
/r/: C-Sub /r/ → [ɾ] > [l], [+vibtr] → [+vibtp] > [-lat][+vibtr] →
[+lat][- vib]
[ɹ] [+cs][+vibtr][Cor] →
[-cs][-vib][Cor][Dor][Lab]
3-year-olds: WM position
Again, patterns are presented in terms of prominence. The most common WM consonant
mismatch in 3-year-olds was consonant substitution in left-, centre- and final-prominent words
followed by consonant deletion. In the higher match categories (over 70%), consonant epenthesis
occurred in both left- and centre-prominent words and consonant lengthening and syllable
unstressed syllables affecting both consonants and syllables. A smaller number of phonemes in
the sample were assessed in right-prominent words, and consequently the number of mismatches
was also small. Similar mismatch patterns were noticed word medially as word initially, i.e. (a)
single and multiple feature substitutions for one segment; and (b) more feature changes in later-
developing phonemes. For example, single feature changes included simple devoicing as when
[ɣ] was produced as [x], or a single place change from [Dorsal] to [Labial] when [ɣ] was
55
produced as [β]. A more complex examples is the case of /l/ with its features
Dorsal]. In general, the features [+/-lateral] and [+/-vibrant] had frequent mismatches across
prominence types. That is, /l/ lost [+lateral] by becoming [j] in the previous example, but in a
different token [ð] went from [-lateral] to [+lateral] when substituted by [l]. The tap /ɾ/ lost its
vibrant quality when substituted by [-vibrant] [n], but [ð] became /ɾ/ [+vibrant] in a different
example. Thus, sometimes a consonant would appear as a substitution for another consonant but
/ɾ/: C-Sub > C-Epen /ɾ/ → [ð], [ɾˡ] > [+vibtp] → [-vib]; [-lat] → [+lat]
/l/: C-Sub > C-Del /l/ → [j], [ɾ] > [+lat] → [-]; [+lat][-vib] → [-lat][+vib]
[ð]: C-Sub [ð] → [ɾ] > [l] [-vib] → [+vibtp] > [-lat] → [+]
/r/: C-Sub /r/ → [ɾ], [ɾˡ], [ɾʲ] > [+vibtr] → [vibtp] (+lat/-cs)
56
[l] (or [l:]) > [+vibtr][-lat] → [+lat][-vib] ([+long])
[-vib][- ct]
/ɲ/: C-Lg
[ɹ] [+cs][+lat][Cor] →
[-cs][-lat][Cor][Dor][Lab]
[ɣ]: C-Sub > C-Del [ɣ] → [w] > [Dor][-rd] → [Lab][+rd] >
57
3-year-olds: WM Centre-prominent - Unstressed syllables
/p/: C-Sub > Syl-Del /p/ → [t] > [j] [Lab] → [Cor]; [Lab][-son] →
[Cor,-ant][+son]
/f/: C-Sub > C-Epen /f/ → [p], [θ] [+ct] → [-]; [Lab] → [Cor,+ant]
[ɣ]: C-Del
/r/: C-Sub /r/ → [ɾ] > [ð] [+vibtr] → [+vibtp]; [+vibtr] → [+vib]
58
3-year-olds: WM Right-prominent - Stressed syllables
Consonant mismatch patterns that appeared in the 4-year-old sample are detailed in this section.
4-year-olds: WI position
Mismatches occurred only in the higher match categories (>70%). Consonant substitutions were
the only pattern observed in this near-mastery category of singletons in WI position across
prominences. The substituted phonemes were for /r/ in stressed and unstressed syllables of left-
and right-prominence respectively, and labial and dorsal voiced stops /b, g/ in unstressed
syllables of centre-prominent words. More feature changes happened apparently in relation to the
complexity of the phoneme, such as a simultaneous changes in manner, place and voice from
[-voiced][Coronal-Dorsal] [ʝ̥]; or, very simple changes in earlier acquired sounds such as manner
(e.g. [+/-nasal] /b/ produced as [m]), or voicing only (e.g. /b/ produced as [p]).
59
Mismatches and dialectal variants: Age 4 (Table 13)
60
4-year-olds: WI Left-prominent - Stressed Feature changes in substitutions
[-vib][-son][Cor,-ant]
[g] [+vibtr][+son][+ct][Cor] →
[-vib][-son][-ct][Dor]
[ʝ̥] [+cs][+vibtr][+vd][Cor] →
[-cs][-vib][-vd][Cor-Dor]
/b/: C-Sub /b/ → [m] > [p] [-nas] → [+] > [+vd] → [-]
4-year-olds: WM position
Consonant substitution was the common mismatch pattern for near-mastery singletons in WM
position of left-, centre- and final-prominent words occurring in un/stressed syllables. Consonant
deletion also occurred for [-consonantal] [ɣ] in both left- and centre-prominent words in
unstressed and stressed syllables respectively. Nasal /m/ was deleted in WM unstressed syllables
of centre-prominent words. Lastly, consonant epenthesis occurred for these same two phonemes
61
/m/ and [ɣ]; /m/ was epenthesized in the same position in which it was deleted, but [ɣ] was
occurred in unstressed syllables for both left- and centre-prominent words when compared to the
stressed syllables of centre- and right-prominence, suggesting that word stress is relevant in
62
Mismatches and dialectal variants: Age 5 (Table 14)
Consonant mismatch patterns that appeared in the 5-year-old sample are detailed below.
63
Mismatches and dialectal variants: Age 5 (Table 14)
64
5-year-olds: WI position
Far fewer consonant mismatches were observed in 5-year-olds when compared with the younger
groups. WI mismatches occurred in the high match categories (>70%) in unstressed syllables of
centre- and right-prominent words. Similar to ages 3 and 4, consonant substitution was the most
common mismatch. Later-acquired phonemes were affected: /g/ and /r/. These phonemes often
showed simple laryngeal or manner substitutions (e.g. /r/ produced as [ɾ] or [r̥], losing one
feature either [+trill] or [+voiced]). However, broader substitutions of many features were also
deletion was only observed in unstressed syllables of centre-prominent words for /g/ and *[h].
N/A
*[h]: C-Del
/r/: C-Sub /r/ → [ɾ] > [r̥] [+vibtr] → [+vibtp] > [+vd] → [-vd],
[-son][-ct][-vibtr][-vd][Dor]
[ɬ] [+son][+vibtr][-lat][+vd] →
[-son][+lat][-vd]
65
5-year-olds: WM position
stressed syllables of centre-prominent words. No mismatches were found in left- and right-
prominent words, or unstressed syllables of centre-prominent words. The substitutions for [ɣ]
involved place only (e.g. [Dorsal] to [Labial] for [ɣ] produced as [β]), or place and manner
N/A
[ʀʷ] [-cs][Dor,+high][-rd] →
[+cs][+vibtr][Dor,+low][+rd]
DISCUSSION
This paper analyzed single word productions of preschool speakers of Granada Spanish ages 3 to
5 in order to determine the order of acquisition of singletons when influenced by word stress and
children got older. Regarding stress, better consonant accuracy in stressed syllables was expected
66
when compared with unstressed syllables independent of word position. In terms of consonant
matching, we expected the WI and WM data to match adult targets of Granada Spanish and other
reported dialects of Spanish. Lastly, mismatch patterns of Spanish were expected to correspond
trisyllabic, 10 tetrasyllabic and 1 word with five syllables. For details regarding the list of words
used, please refer to Appendix 2. Words were grouped by WI and WM onset position of
singletons, and sub-grouped by left, centre and right stress prominence. Each subgroup was
further labeled as “stressed syllable” or “unstressed syllable” depending on the stress of the
syllable in which the target phoneme was included. Phonemes in WM onset position of centre-
prominent words were the only ones to occur in both stressed and unstressed syllables as
mentioned throughout the results. Unfortunately, our sample did not include a sufficient number
of tokens for every phoneme by position. For those marked with an asterisk, data needs to be
interpreted cautiously. Also, some phonemes were evaluated for one age group or one position
but not for another. As a result, it was not possible to establish whether a phoneme was acquired
comparatively among those specific ages or groups. If a phoneme is not noted in the tables, it
was not elicited, either due to not being included in the elicitation word list, or because the given
phoneme does not happen in that position in Granada Spanish, e.g., /p/ does not occur word
finally in Spanish. Data showed that the trilled /r/ at age 4 was mastered word medially in
unstressed syllables of centre-prominent words, but because no other /r/ was elicited in stressed
syllables in that same context, we could not determine whether the production of /r/ in stressed
and unstressed syllables would have been acquired at different times. Despite these limitations,
the results contribute new information to the literature on Granada Spanish and similar dialects
67
and show similarities with other previous research. In the subsequent discussion, we first discuss
findings of the current study relative to changes in developmental accuracy and compare this
with reports in the literature, and then discuss consonant mismatches in this study, with a second
comparison with the literature. The paper ends with implications for future research and clinical
application.
Children in our sample, aged 3 to 5, showed increasing mastery of singleton production in both
WI and WM positions as they got older (see tables 10 and 11). Phonemes of earlier acquisition
present for [+continuant] and [+approximant], but those categories contain a larger number of
phonemes and vary in adult speech. The last phoneme to be acquired was [+trilled] /r/. Our
sample also did not show mastery of the approximant [ɣ] by age 5.
As mentioned, the words included in this study were categorized into left-, centre- and
right-prominent word positions, and sub-categorized into context of unstressed and stressed
syllables. The data showed overall later acquisition of some phonemes in unstressed syllables of
centre-prominent words.
This specific finding will be explained mainly for age 3 because that is when syllable
stress and prominence were particularly relevant to acquisition. As mentioned before, the
68
younger the children were, the fewer matches they presented when compared with older groups.
The phonemes /t, b, d, x/ were not mastered word initially in unstressed syllables of centre-
prominent words, but were mastered in all other contexts that were elicited. The /p, n, f/ were not
mastered in unstressed syllables of centre-prominent words, but were mastered in all other
elicited contexts. The nasal /m/ was also not mastered in both unstressed environments, but was
mastered elsewhere as elicited. As had been predicted, the study showed higher consonant
accuracy in stressed than unstressed syllables (Bernhardt & Stemberger 1998; Chávez-Peón et al.
Phonemes of higher complexity were present in the 4-year-old sample compared with the
3-year-old sample, where there was additionally, more variation for these targets. For example,
[ɾ] was below mastery at age 3 for the three elicited WM prominences that included both stressed
and unstressed syllables but all were mastered by age 4. At age 3, the trill showed still-
developing or emerging levels across positions, syllable stress and prominence. At age 4, trill
showed near mastery for both stressed and unstressed syllables and more surprisingly, mastery
only in WI and WM unstressed syllables of centre-prominent words. For the 4-year-olds, the
approximant variant [β] was the only phoneme not mastered in stressed syllables of centre-
prominent words, but was mastered elsewhere. These details raise another hypothesis in need of
further research, i.e. that some phonemes of higher complexity may take longer to develop but
because they do, are then less prone to vulnerability in the by then more well-developed initial
unstressed syllable.
69
Impact of stress of the syllable and length of the word
The most common structure that children hear in Spanish is disyllabic words with left
monosyllables (S) (Lleó, 2006). Trochees are likely easier for the children to produce because of
how frequently they hear them, their short length (only two syllables - one foot) and their initial
stressed syllable. We mentioned above that most frequently, later acquired phonemes in our data
were found in unstressed syllables of centre-prominent words. In order for a word to have centre
prominence, it needs to have a minimum of three syllables, which adds complexity. First, word
length is increased (trisyllabic), and second, the stress is no longer at the beginning of the word.
If it is considered that a foot is made of a maximum of two syllables the vast majority of the time
in Spanish (Lleó, 2006)— one strong or stressed, and one weak or unstressed — an additional
(or extrametrical) syllable then is posited to be located on the left or the right edge of the foot.
The phonological system of young children may allow for one foot per word at a specific time,
so if they encounter a longer word, they must repair it by deleting a syllable or changing the
stress of the word (Bernhardt & Stemberger, 2000). In our data, we observed most syllable
deletions in centre-prominent words, which as explained above, are longer than the easier to
produce disyllabic ones. A 3-year-old in our sample said ['fo.no] instead of /te.'le.fo.no/
'telephone'. In this example, he changed the stress from the second to the third syllable, deleted
the first and second syllables, and built a disyllabic left-prominent word. Another 3-year-old
to /ˈɲeka/ 'doll'. The 4- and 5-year-olds did not show as many syllable deletions as the 3-year-
70
olds, because at that age their system perhaps did allow extrametrical syllables. In addition to
syllable deletion in multisyllabic words, it is important to look at how consonants were affected
The trisyllabic words elicited for this study included two left-prominent ones (i.e.
lámpara, pájaro), one right-prominent (e.g. pantalón) and the rest, centre-prominent (e.g.
were deleted at ages 3 and 4. This supports the hypothesis that not only stress but also word
length affects production accuracy. At age 5, however, one WI syllable was deleted in a centre-
prominent word and another WI syllable was deleted in a left-prominent word. If not only stress
and word length are important, maybe other factors are involved, e.g. consonant sequences.
Consonant sequences (e.g. labial-dorsal-coronal in the word pájaro ['pa.xa.ɾo] 'bird') are
a factor to consider in multisyllabic words because they might have greater complexity than that
allowed by the phonological system of a child at a given time (Bernhardt & Stemberger, 1998). It
would be necessary to analyse the productions of each child separately in order to determine if,
in these data, consonant sequences were an influential factor. Although this is not the focus of
this paper, we do propose further research on this topic. It could be one reason for certain
Lastly, word length is influenced by the complexity of its syllables based on the
consonant-vowel strings; either syllables that are formed as single vowels (e.g. uvas /'u.βas/
/pri.ma.'ve.ɾa/ 'spring'). For example, a 3-year-old said [so.ˈβe.ɾo] instead of /som.ˈbre.ro/ 'hat'.
He kept the number of syllables but reduced the word structure CVC.CCV.CV to CV.CV.CV,
71
In summary, acquisition of singletons was affected by several factors: word length (number of
syllables), word stress (left-, centre-, right-prominence), position of the consonant in the word
(WI, WM, WF, sequences), and type of stress in which the singleton is located (un/stressed
syllable).
The results of this study were generally similar to those described in the literature (Jiménez,
1987; Bosch, 2004; Vivar & León, 2009). Singleton voiceless stops ([-sonorant][-continuant][-
voiced]) were mastered at age 3, as indicated in our relative mastery tables 10 and 11, as well as
in Jiménez (1987) and Bosch (2004). A slight variation occurred with /t, p/ in WI and WM
prominent words (e.g. deletion of /p/: Europa ˈEuropeˈ /eu.ˈɾo.pa/ produced as [eu̯ .ˈo.ta]),
suggesting some effect of prominence context. The voiced bilabial stop /b/ was acquired at age 3
in Jiménez (1987), Bosch (2004) and Vivar and León (2009), but in our sample, /b/ only showed
‘dancing’ /bai.ˈlan̪ .do/ produced as [a̯ .ˈlan̪ .do]), i.e. perhaps stress was relevant in the lower
Mastery of nasals ([+nasal]) /m, n, ɲ/ was achieved at age 3 according to Jiménez (1987),
Bosch (2004) and Vivar and León (2009). Our data concurred approximately, in that /m, n/ were
mastered in stressed syllables of other prominences. However, stress effects were noted: /m/ in
mastery, examples being: (a) WI /m/: muñeca /mũ.ˈɲe.ka/ ‘doll’ produced as [ũ.ˈɲe.ka]; (b) WM
72
/m/: hipopótamo /i.po.ˈpo.ta.mo/ ‘hippopotamus’ produced as [e.po.ˈpo.ta.no]). The WM /n/
showed similar patterns, being a match word medially in centre-prominent words in stressed
syllables chimenea /ʧi.mẽ.ˈne.a/ ‘chimney’ but in unstressed syllables showing substitutions, e.g.
In this study, there was a wider variety of sounds acquired by age within the different
word stress categories when compared to the single category used by Jiménez (1987), especially
for fricatives [+continuant]. We concurred more with Vivar and León (2009) (i.e. /f, x/), and with
Bosch (2004) (i.e. /θ/). Cataño et al. (2009) considered /θ/ as a unique phoneme to Peninsular
Spanish which led them to exclude it from their Spanish phonemic inventory; yet they mentioned
that it was presented by some of their participants as an allophone [θ]. The phoneme /f/ was
mastered at age 3 in WI and WM stressed syllables of left- and centre-prominent words but
nearly mastered in unstressed syllables of centre-prominent words in our study. Jiménez (1987)
and Bosch (2004) showed acquisition of /f/ by age 4. In our study, /x/ was acquired at this same
age except word initially in unstressed syllables of centre-prominent words where it was still
developing. In comparison, /x/ was acquired by age 3 in Bosch (2004) and Vivar and León
(2009). The /s/ showed slightly higher developing levels in our study at age 3 than in previous
The voiced approximant allophone [ʝ] was acquired at age 3 in the present study and
others (Jiménez, 1987; Vivar & León, 2009). Bosch (2004), who did not report [ʝ] but /ʎ/ found
that it was acquired at age 4 - (/ʎ/ may be treated as an approximant). In our results, WI /d, g/
(1987), but Vivar and León (2009) observed higher mastery. Vivar and León did not actually
report on the voiced approximant allophones [ð, ɣ]. An interesting distribution of these
73
allophones can be observed across stages in our sample at age 3 where they spread over different
developing categories (near-mastery, still developing and emerging). By age 4, Jiménez (1987)
considered /d, g/ as acquired; [ð] was mastered in our sample, and [ɣ] was nearly-mastered. One
could argue that Jiménez did not elicit the approximant allophones because his population was
from California in which the children could have been more influenced by the English
production of the voiced stops. Acevedo (1993) reported on the allophones [β, ð, ɣ]; [β] and [ð]
were acquired by age 3, and [ɣ] by age 5. Vivar and León’s (2009) data showed acquisition of /g/
by age 5. In the current study, /g/ remained at near-mastery level in WI position in unstressed
syllables of centre-prominent words, and [ɣ] also, but word medially. Acevedo (1993) reported
that the only two phonemes in her data that did not reach mastery by age 5 were /g/ and /r/. Thus,
Variability in the allophonic differences and the phonemes they approximate might
suggest the difficulty it takes for children to organize these variations, and thus more time could
be required to master them (Cataño et al., 2009). The frequency of the phonemes in a language is
said not to be sufficient reason to explain the variability of acquisition; rather, it could be
explained by the role phonemes and allophones play to differentiate one word from another. For
example, in the case of fuego ‘fire’, the allophonic productions of [ˈfue.go] and [ˈfue.ɣo] do not
mark a semantic contrast. Therefore, the speakers of the language would not have a crucial need
to develop the production of both phoneme and allophone, as happens with other languages
(Cataño et al., 2009). Maybe this is why in the present research, we did not find as many tokens
for some allophones in different positions as indicated with the asterisk in the tables 7-11, and
why the allophone [ɣ] took longer to be acquired and even showed a regression at age 5.
74
Jiménez (1987), Bosch (2004), Vivar and León (2009) and the present study concur
regarding acquisition of the tap at age 4. All but Bosch agreed on the trill being mastered at age 5
(she had found mastery of trill after age 6). We found that in WI position of unstressed syllables
of right-prominent words, /r/ remained at near-mastery level. Lastly, the affricate /ʧ/ was
acquired at age 4 in all studies, earlier than in Jiménez who reported it as developing at age 4
without further detail of acquisition. Interestingly, none of these studies reported the affricate /ʤ/
which was not included in the Spanish inventory presented by Cataño et al. (2009). In our
sample, /ʤ/ was mastered at age 3 word initially in un/stressed syllables of centre- and left-
Overall, age of acquisition across the reviewed studies (Jiménez, 1987; Bosch, 2004;
Acevedo, 1993; Vivar & León, 2009) was consistent with ours regarding liquids (lateral, tap and
trill); for obstruents (stops, fricatives, affricates) and nasals, more variation was found. It may be
relevant that the obstruents and nasals have more diverse places of articulation from bilabial,
labiodental, coronal [+anterior], and dorsal, and their laryngeal features are evenly spread [+/-
voiced][+/-spread glottis] whereas the liquids are within the same categories for manner, place
and voicing. Thus, potential for variation across studies may be higher for the more diverse
category.
position, and thus position of the given onset did not appear to be a main affecting factor for the
Overall, this discussion compared the acquisition of phonemes of the Granada Spanish
dialect and others which supports our prediction of how our participants would match the adult
75
targets of these Spanish dialects. Although our data coincided generally with the existing
literature, our data also showed that some phonemes remained below mastery level when they
The current study: General patterns relative to the stress of the syllable
Syllable structure may be modified when phonemes are difficult to produce (Bosch, 2004), either
2000; Bernhardt et al., 2015). We observed that syllables were deleted more often in
multisyllabic words. Higher cognitive demands at the level of the syllable may consequently
compromise lower levels of the phonological system at a phonemic and feature level in a
Syllable deletions were observed mostly in centre-prominent words. They happened more
frequently at age 3 (see Tables 12-14). Word-initial unstressed syllables were deleted only in
centre-prominent words, while at age 5, very few were deleted in either left- or centre-prominent
words. At age 3, WM pre-stress syllables were deleted; at age 4, pre- and post-stress deletion
was found; and at age 5, only post-stress deletion was observed. Consonant deletions were more
commonly found in stressed and unstressed syllables of different prominence for age 3. This
could be an indicator of the lack of maturity of singleton production at this age, which is further
supported by the large number of consonant substitutions. The 4-year-olds had much fewer
consonant deletions; these occurred in unstressed syllables in WI and WM positions, and stressed
76
syllables in WM position in left- and centre-prominent words. The 5-year-olds only showed two
age 3 (WI Centre - 9, WI right - 1, [WI total: 10] versus WM - 5). This pattern was not observed
at age 4 (WI Centre - 2 versus WM Centre - 2) nor age 5 (WI Centre - 2 versus WM Centre - 0).
3-year-olds
The most common mismatches found at age 3 were consonant substitutions. In WI position in
stressed syllables of left-prominent words, the substitutions generally showed a change in one
feature. For example, /s/ appeared as [t], with a change in manner of articulation from
[+continuant] to [-continuant]. Also, /t͡ ʃ/ appeared both as [ʃ] (acceptable variant), changing
either manner of articulation losing the [-continuant] quality, and as [t], losing [-continuant].
Some changes were observed for place of articulation, such as /s/ appearing as [-anterior]
[-grooved] [ɕ]. In WM position, it was more common to observe changes in two features
Also, for more complex phonemes such as the rhotics, a wide range of substitutions was found.
Trilled /r/ was produced as [ð], [ɾˡ], [l], [l:], [n] or [β]. Its core feature [+vibrant] was lost
everywhere except for [ɾˡ]. Phonemes that were at a still-developing or emerging level showed
even less accuracy. For example, /x/ surfaced as [t], [f], [j], [kwx]. The [t] substitution shows
changes in manner [-continuant] and place [+coronal][+anterior], but the [j], manner, place and
77
laryngeal features. Lastly for [kwx], although the /x/ is kept, manner and place features are added
with [kw].
It is possible that disyllabic left-prominent words, which appear more often in the
language they hear and is the most frequent structure used by very young children, provided a
well-practiced frame for the children to produce phonemes, and hence they presented fewer
mismatches in such words, as compared with longer words without initial stress.
4-year-olds
Substitution patterns were similar to those of the 3-year-olds but much less frequent. For
complex phonemes, there was much more variability in the features that changed for the target
phoneme as in the example of /r/ appearing as [d͡ ʒ], [g] or [ʝ̥], but more simple phonemes
undergoing one change as in /b/ appearing as [m] or [p], i.e. being either nasalized or devoiced.
Mismatches occurred showing differences based more on the phoneme than on the position in
5-year-olds
Mismatches at age 5 were minimal. Changes in place, manner and/or laryngeal features
persisted, probably due to the level of complexity of the phonemes as mentioned before.
Devoicing was observed in all substitutions of /r/ within other mismatch patterns. Particularly,
rounding was observed for [ɣ], appearing as [β] but also as [ʀʷ], [w], while [ɣ] had only been
78
Mismatch comparison with previous research on Spanish
Mismatch patterns presented by the children of our sample are compared with the phonological
evaluation developed by Bosch (2004) because it is one of the more complete analyses for
European Spanish. The important variation is that of codas which are present in Castilian
Spanish (Bosch's study) and some of the other phenomena of the Granada dialect. We will not
compare WF phonemes to the Bosch study because it is not the focus of this paper because of the
Bosch (2004) described systemic 'processes' that commonly appear at certain age periods.
Bosch (2004) determined the following to be frequent for 3-year-olds: absence of trill, stridency
insertion for /θ/, realized as [s, f], fronting of /s/, lateralization of [ð] and gliding of /ʎ/ to [j]
(Bosch, 2004. p. 69, paraphrased and translated). Our sample also lacked trilled /r/. As for /s/,
more backing occurred rather than fronting in our cohort (/s/ as [ɕ], [s̞ ], [ʃ]). Instead of
lateralization of [ð], we found substitution of [ð] for /l/. We did not find gliding of /ʎ/ because /ʎ/
is not a phoneme in Granada Spanish. But we did find gliding in other contexts: e.g. /θ/ ([h]), /l/
Bosch (2004) mentioned stopping of fricatives as an risk indicator for PPD. In our TD
sample of 3-year-olds, we found this pattern rarely, and only in WI and WM unstressed syllables
of centre-prominent words e.g. /x/ as [t], /f/ as [p]. It would be interesting to observe if the tokens
79
elicited by Bosch included fricatives in words of different prominence in order to support her
Bosch (2004) considered that deletion of WI consonants and unstressed syllables was a
risk factor for PPD at this age. While she determined the stress of the syllables that were deleted,
she did not mention the type of stress in syllables in which consonant deletions was common.
Although consonant substitutions were the most common mismatch in the current study, there
there were: (a) syllable deletions affecting [-continuant] /b, t/ and [+continuant] /θ, s/; and (b)
consonant deletion of [-continuant] /b, m/ and affricate [+continuant][-continuant] /t͡ ʃ/. Word
medially, there were consonant deletions for [+approximant] [ɣ] and /l/ in stressed syllables of
left-prominent words, and syllable deletions for [+lateral] /l/ in stressed syllables of centre-
prominent words. Further, in unstressed syllables of centre-prominent words, there were: (a)
syllable deletions of [+sonorant] /p/, /n/; and (b) consonant deletions of [+approximant] [ɣ], /l/.
Finally, in right-prominent words, WM [+lateral] /l/ was deleted. Thus, a number of deletions
still occurred in TD 3-year-olds, but these appeared to relate to the segment, the word position
(more WM) and the word stress/length. Deletion per se may not be a risk factor, but where
deletion occurs.
Bosch (2004) noted common mismatch patterns at age 4 to be similar to those of 3-year-olds:
absence of trill, stridency of /θ/ (produced as [s, f]), fronting of /s/, and gliding of /ʎ/, produced
as [j] (Bosch, 2004. p.70, paraphrased and translated). In contrast, we found /r/ to be at a mastery
80
and near-mastery level in WI and WM position. We found stridency of /θ/ (produced as [s]) in
centre-prominent words /s/ as [θ] (both acceptable variants in Granada). Fronting was found for
/s/ ([s̪ ]), [ɣ] ([β]) and /g/ ([t]). Fronting was determined as a risk indicator for Bosch (2004) at
this age, among other processes that were not shown in our sample. Interestingly, lateralization
of all trills was another risk indicator. We did find this process but accompanied by others /r/ as
Similar to age 3, Bosch (2004) considered that deletion of WI consonants and unstressed
syllables was a risk factor at this age. Interestingly, /m/ deletion did occur in the current study in
centre-prominent words, i.e. rare, but occurring in longer words and thus not a risk factor.
The common mismatch processes expected at age 5 were absence of trill, gliding of /ʎ/ ([j]), and
stridency of /θ/ ([s]) (Bosch, 2004. p.71, paraphrased and translated). The trill /r/ was mastered in
the current sample in all positions in this age group except for WI unstressed syllables of right-
prominent words in which it was near mastery. We observed gliding for /g/ ([j]) of WI unstressed
words.
The comparison of the present data and Bosch’s (2004) for mismatch patterns suggest
that acquisition of Granada Spanish might have systemic and structural processes particular to
81
the Granada dialect. Further investigation is suggested, particularly with a range of words in
Sample size
Our sample size (30 children) is small and thus may not be an actual representation of the
general population. Our intention was to obtain criterion reference data for acquisition of
phonemes when influenced by stress and prominence. Given that the number of children who
participated was low, the results are preliminary only. Further research with larger samples is
needed.
Not all the phonemes were measurable across prominence and word position because of
differences in the Spanish language and the restrictions on words known by children (and in a
number that they can tolerate to say during testing). Consequently, it was difficult to determine
exact effects of consonant, stress, word position and prominence. More data would allow a
statistical comparison and stronger evidence. Performing a study that analyzed accuracy of
consonant matches based on frequency of the phonemes in the language would also be useful.
Some phonemes may be acquired earlier than later when its frequency in the language is higher
82
Clinical Implications of the Current Study
As presented here, the results that compare the phonemes across prominence and syllable stress
allow us to observe that phonemes may behave differently depending not only on the position
that they occupy in the word, but also on the stress of the syllable that they occupy, the
prominence of the word, and word length. Clinically, speech-language pathologists need to
consider such results not only to compare the child’s phonological acquisition with the norm, but
to assess and treat phonemes in as varied contexts as allowed in that specific language. That is, if
/r/, in Granada Spanish is used WI and WM only, then words with different stress and length
should be used, as well as the /r/ being in stressed and unstressed syllables. Although the focus of
this paper is on word position and stress, we are not excluding the importance of word shape
CONCLUSION
The present study makes a contribution to data on consonant acquisition of Granada Spanish in
relation to word stress and word position, showing the relevance of dialect, word position, word
length, prominence patterns and stress on match and mismatch patterns for consonants. Further
investigations with both this and other dialects, larger sample sizes and tokens per phoneme are
warranted.
83
ACKNOWLEDGMENTS
I would like to thank my advisor, Prof. May B. Bernhardt, for her patience and invaluable
guidance during this research and graduating project, and my education in general. I am thankful
to the researchers involved in the investigation this paper is based on, and to the children and
families who participated in the original study. Thank you also to the Social Sciences and
Humanities Research Council of Canada grant 410-2009-0348 and the Junta de Andalucía in
Spain, Grupo Hum-605, Logopedia Experimental y Aplicada [Experimental and applied speech-
language pathology].
84
APPENDIX 1. Nonlinear Scan Analysis form for Spanish
[+vibrant] ([+rhotic]) r ɾ
[+vibrant trill] [+vibrant tap]
(or [+multiple vibrant] [+simple vibrant])
[+lateral] l
[+labiodental] f
85
APPENDIX 2
WORD INITIAL
Word Position Word Stress Phoneme appears in:
Phoneme Orthography Phonetic (prominence)
Transcription: S: stressed
Adult WI: word initial L: left syllable
WM: word C: centre U: unstressed
medial R: right syllable
WF: word final
bañera ba.ˈɲe.ɾa WI C U
baño ˈba.ɲo WI L S
barco ˈbaɾ.ko WI L S
boca ˈbo.ca WI L S
veinte ˈbein.te WI L S
k casa ˈka.sa WI L S
caballo ka.ˈβa.{ʝ/jo WI C U
cocodrilo ko.ko.ˈ{d/ð}ɾi.lo WI C U
conejo ko.ˈne.{x/h}o WI C U
͡ tʃ chimenea ʧi.mẽ.ˈne.a WI C U
chocolate ʧo.ko.ˈla.te WI C U
dinosaurio di.no.ˈsau.ɾjo WI C U
f foto ˈfo.to WI L S
ˈfut.bol/ WI L S
fútbol
ˈfu.{b/β}(b)o/{ɔ}(l)
gorra ˈgo.ra WI L S
86
Phoneme Orthography Adult Word Position Word Stress Stressed/Unstressed
guitarra gi.ˈta.ra WI C U
{x/h}a.ˈmõn / WI R U
x/h jamón
ha.ˈmõ / xa.ˈmɔ
l lámpara ˈlam.pa.ɾa WI L S
leche ˈle.{ʧ/ʃ}e WI L S
ʤ llave ˈʤa.βe WI L S
m martillo ma{ɾ.ˈt/{ˈt.t}i.ʝo WI C U
mesa ˈme.sa WI L S
muñeca mũ.ˈɲe.ka WI C U
nieve ˈnie.βe WI L S
noche ˈno.{ʧ/ʃ}e WI L S
p pájaro ˈpa.xa.ɾo WI L S
pan ˈpan WI L S
pantalón pan.ta.ˈl{on/ õ} WI R U
papá pa.ˈpa WI R U
Paula ˈpau.la WI L S
pelo ˈpe.lo WI L S
87
perro ˈpe.ro WI L S
r ratón ra.ˈt{on/ õ} WI R U
regalo re.ˈɣa.lo WI C U
rojo ˈro.{xh}o WI L S
silla ˈsi.ʝa WI L S
θa.na.ˈo.ɾja / WI C U
zanahoria
sa.na.ˈo.ɾja
sa.ˈpa.to / WI C U
zapato
θa.ˈpa.to
sombrero som.ˈbɾe.ɾo WI C U
t techo ˈte.{ʧ/ʃ}o WI L S
teléfono te.ˈle.fo.no WI C U
toca ˈto.ka WI L S
tortuga to{ɾ.ˈtu/ˈt.t}.ɣa WI C U
88
APPENDIX 2. List of Words (Continued)
WORD MEDIAL
ka.ˈ{b/β}a.{ʝ/j}o
b/β caballo WM C S
etc
ˈfut.bol/ WM L U post-stressed
fútbol
ˈfu.{b/β}(b)o/{ɔ}(l)
primavera pɾi.ma.ˈβe.ɾa WM C S
pescado peʰ.ˈka.{ð/d/ø}o WM C S
89
Phoneme Orthography Adult Word Position Word Stress Stressed/Unstressed
d/ð/ø pescado peʰ.ˈka.{d/ð/ø}o WM C U post
dragón {d/ð}ɾa.ˈɣ{on/ õ} WM R S
regalo re.ˈɣa.lo WM C S
chocolate ʧo.ko.ˈla.te WM C S
escalera eʰ.ka.ˈle.ɾa WM C S
pantalón pantaˈl{on/õ} WM R S
90
pelo ˈpe.lo WM L U post
teléfono te.ˈle.fo.no WM C S
{x/h}a.ˈmõn /
jamón WM R S
ha.ˈmõ / xa.ˈmɔ
conejo ko.ˈne.{x/h}o WM C S
θa.na.ˈo.ɾja /
zanahoria WM C U pre
sa.na.ˈo.ɾja
91
Phoneme Orthography Adult Word Position Word Stress Stressed/Unstressed
ɲ bañera {b/β}a.ˈɲe.ɾa WM C S
muñeca mũ.ˈɲe.ka WM C S
hipopótamo i.po.ˈpo.ta.mo WM C S
papá pa.ˈpa WM R S
zapato sa.ˈpa.to/θa.ˈpa.to WM C S
Europa {eu/u}.ˈɾo.pa WM C S
jirafa {x/h}i.ˈɾa.fa WM C S
92
Phoneme Orthography Adult Word Position Word Stress Stressed/Unstressed
s azul a.ˈsu{l/ø/ɾ} WM R S
dinosaurio di.no.ˈsau.ɾjo WM C S
guitarra {g/ɣ}i.ˈta.ra WM C S
ratón ra.ˈt{on/õ} WM R S
93
APPENDIX 2 (Continued)
WORD FINAL (optional in the Granada dialect except perhaps /n/ in monosyllables)
Phoneme Orthography Phonetic Word Position Word Stress Phoneme appears in:
Transcription: (prominence)
Adult S: stressed
WI: word initial L: left syllable
WM: word medial C: centre U: unstressed
WF: word final R: right syllable
h bloques ˈblo.kɛh WF L U
lápiz ˈla.pih WF R S
reloj re.ˈloh WF R S
x reloj re.ˈlox WF R S
l azul a.ˈsul WF R S
jamón {x/h}a.ˈmon WF R S
pan ˈpan WF L S
pantalón pan.ta.ˈlon WF R S
ratón ra.ˈton WF R S
ŋ pan ˈpaŋ WF L S
θ cruz ˈkɾuθ WF L S
lápiz ˈlapiθ WF L U
luz ˈluθ WF L S
nariz na.ˈɾiθ WF R S
pez ˈpeθ WF L S
s bloques ˈblo.kes WF L U
94
Phoneme Orthography Adult Word Position Word Stress Stressed/Unstressed
s brazos ˈbɾa.sos WF L U
chocolates ʧo.ko.ˈla.tes WF C U
cruz ˈkɾus WF L S
dos ˈdos WF L S
escaleras e{s/ʰ}.ka.ˈle.ɾas WF C U
luz ˈlus WF L S
frutas ˈfɾu.tas WF L U
globos ˈ{g/ɣ}lo.{b/β}os WF L U
gracias ˈɡɾa.{s/θ}jas WF L U
lápiz ˈla.pis WF L U
pez ˈpes WF L S
tres ˈtɾes WF L S
uvas ˈu.βas̙ WF L U
zanahorias {s/θ}a.na.ˈo.ɾjas WF C U
zapatos {s/θ}a.ˈpa.tos WF C U
95
APPENDIX 3. Number of tokens per phoneme
b 28 31 39 WI L S
10 15 12 WI C U
β 8 8 4 WI L S
3 4 6 WI C U
k 9 10 11 WI L S
27 30 33 WI C U
͡ tʃ 26 30 33 WI C U
d 16 18 22 WI L S
9 9 11 WI C U
ð 3 2 - WI L S
f 18 20 22 WI L S
g 16 19 20 WI L S
7 8 11 WI C U
ɣ 2 1 2 WI L S
1 2 - WI C U
x 7 9 11 WI L S
18 19 18 WI C U
9 5 8 WI R S
h 2 1 - WI L S
96
- 1 3 WI C U
5 - 2 WI R S
l 36 40 43 WI L S
ʝ 3 2 1 WI L S
1 2 - WI C U
d͡ ʒ 6 8 10 WI L S
8 8 11 WI C U
m 9 10 11 WI L S
18 20 22 WI C U
n 9 10 11 WI L S
9 10 11 WI R S
p 64 89 86 WI L S
9 10 12 WI C U
18 21 22 WI R S
r 9 10 10 WI L S
9 10 11 WI C U
22 30 29 WI R S
s 18 20 22 WI L S
27 30 22 WI C U
θ 14 19 28 WI C U
t 18 20 21 WI L S
18 20 22 WI C U
97
WORD INITIAL - NUMBER OF TOKENS PER PHONEME
β 61 61 68 WM L U post
18 20 22 WM C S
k 27 30 32 WM L U post
18 10 11 WM C S
9 10 11 WM C U post
18 20 22 WM C U pre
͡ tʃ 31 39 44 WM L U post
ð 9 8 11 WM L U post
4 4 10 WM C U post
f 9 10 11 WM C S
9 10 11 WM C U post
9 10 11 WM C U post med
ɣ 23 20 20 WM L U post
17 20 21 WM C S
9 10 11 WM C U post
9 10 11 WM R S
x 26 26 31 WM L U post
9 - 11 WM C U post
- 10 - WM C U post med
h 1 4 1 WM L U post
l 39 50 50 WM L U post
98
Phoneme Age 3 Age 4 Age 5 Word Position Word Stress Stressed/
(prominence) Unstressed
l 38 40 44 WM C S
- 20 - WM C U post
9 10 22 WM C U pre
22 30 29 WM R S
ʝ 24 30 30 WM L U post
17 20 22 WM C U post
d͡ ʒ 2 - 3 WM L U post
m 9 10 11 WM L U post
9 10 11 WM C U post
18 20 22 WM C U pre
9 10 10 WM R S
n 11 19 18 WM L U post
18 20 24 WM C S
19 20 22 WM C U post
18 19 20 WM C U pre
ɲ 9 10 11 WM L U post
13 19 18 WM C S
nn 15 5 10 WM L U post
p 9 10 10 WM L U post
21 31 29 WM C S
9 10 11 WM C U post
10 10 11 WM C U pre
9 10 11 WM R S
ɾ 35 39 44 WM L U post
28 30 33 WM C S
99
Phoneme Age 3 Age 4 Age 5 Word Position Word Stress Stressed/
(prominence) Unstressed
ɾ 31 39 41 WM C U post
9 10 11 WM R S
r 18 20 22 WM L U post
8 10 10 WM C U post
s 44 40 46 WM L U post
9 9 11 WM C S
8 10 11 WM C U post
6 1 1 WM R S
θ 8 10 19 WM L U post
2 9 10 WM R S
ʃ 4 1 - WM L U post
t 27 30 33 WM L U post
8 20 10 WM C S
31 29 39 WM C U post
9 10 11 WM C U post med
9 10 11 WM R S
100
WORD FINAL - NUMBER OF TOKENS PER PHONEME
ð 2 - 1 WF L S
h 1 1 - WF L U
- - 2 WF R S
x 2 5 4 WF R S
l 3 3 7 WF L U
4 9 10 WF R S
n - 6 - WF L U
- 7 5 WF L S
24 32 31 WF R S
ŋ - 4 5 WF L S
- - 1 WF R S
ɾ 3 13 11 WF L S
s 13 19 20 WF L U
17 12 14 WF L S
1 1 2 WF C U
θ 1 10 11 WF L S
- 3 4 WF L U
- 2 2 WF R S
101
REFERENCES
Aguinaga, G., Armentia, M. L., Fraile, A., Olangua, P. & Uriz, N. (2004). Prueba de Lenguaje
Oral, Navarra Revisada (PLON-R) [Oral Language Test, Navarra Revised (PLON- R)].
Madrid: TEA.
Álvarez Martínez, M. A. (1987). Notas sobre la lengua española hablada en Madrid. [Notes on
spoken Spanish in Madrid]. Revista De Filología Española, 67(3), 331. Retrieved from
http://ezproxy.library.ubc.ca/login?url=http://search.proquest.com.ezproxy.library.ubc.ca/d
ocview/1299259571?accountid=14656
Bernhardt, B., Hanson, R., Perez, D., Ávila, C., Lleó, C., Stemberger, J. P., Carballo, G.,
102
Bosch, L. (1983). El desarrollo fonológico infantil: una prueba para su evaluación [Child
phonological development: a test for its evaluation]. Anuario de psicología, 28(1), 87-114.
Bosch, L. (2004). Evaluación fonológica del habla infantil [Phonological evaluation of child
Calet, N., Mendoza, E., Carballo, G., Fresneda, M. & Muñoz, J. (2010). Test de Comprensión de
Cataño, L., Barlow, J. A., and Moyna, M.I. (2009). A retrospective study of phonetic inventory
Chávez-Peón, M. E., Bernhardt, B. M., Adler-Bock, M., Ávila, C., Carballo, G., Fresneda, D.,
Lleó, Mendoza, E., Pérez, D. & Stemberger, J. P. (2012). A Spanish pilot investigation for
Dubasik, V. L. & Ingram, D. (2013). Comparing phonology of dyads of children with typical
development and protracted development. Clinical Linguistics and Phonetics, 27(9), 705-
719.
Dunn, L., Dunn, L. & Arribas, D. (2006). PPVT-III Peabody: Test de Vocabulario en Imágenes
Ethnologue. (2017). Summary By Language Size. Retrieved Feb 22, 2017 from
https://www.ethnologue.com/statistics/size
103
Grunwell, P. (1985). Phonological assessment of child speech (PACS). Windsor, UK: NFER.
Proceedings Project.
Jiménez, B. C. (1987). Acquisition of Spanish consonants in children aged 3–5 years, 7 months.
Justicia, F., Santiago, J., Palma, A., Huertas, D., & Gutiérrez, N. (1996). La frecuencia silábica
del español escrito por niños: Estudio estadístico [Syllable frequency: A statistical analysis
Kaufman, A. & Kaufman, N. (2009). K-BIT: Test Breve de Inteligencia de Kaufman; Adaptación
Española [K-BIT: Kaufman Brief Intelligence Test. Cordero, A. and Calonge, I. Spanish
Lleó, C. (2006). The acquisition of prosodic word structures in Spanish by monolingual and
Melgar de González, M. (1976). Cómo detectar al niño con problemas de habla. [How to detect
Moreno, A. & Mariño, J. B. (1998). Spanish dialects: Phonetic transcription. In R.H. Mannell & J.
Association, Incorporated.
104
educated Spanish of Granada: linguistic factors].
Pye, C., Ingram, D. & List, H. (1987). A comparison of initial consonant acquisition in English
and Quiché. In K. Nelson & A. van Kleeck (Eds.), Children's language, vol. 6 (pp. 175-
http://childes.psy.cmu.edu/phon/).
Schmitt, L. S., Howard, B. H. & Schmitt, J. F. (1983) Conversational speech sampling in the
Susaníbar, F., Huamaní, O. & Dioses, A. (2013). Adquisición fonética-fonológica [Phonetic and
infantil (CEFI) a una muestra de niños chilenos [Application of a questionnaire for the
105