Review of Salvadorian Spanish in Toronto
|AUTHOR: Hoffman, Michol F.
TITLE: Salvadorian Spanish in Toronto
SUBTITLE: Phonological Variation among Salvadorian Youth in a Multilectal,
SERIES TITLE: Studies in Romance Linguistics 63
Ann Marie Delforge, Department of Spanish and Italian, Montclair State University
This monograph, based on the author's 2004 dissertation, examines the aspiration
and deletion of syllable-final and word-final /s/, the aspiration of word and
syllable-initial /s/ and the velarization and elision of word-final /n/ in the
speech of young Salvadorian immigrants residing in Toronto. Multivariate
analysis is used to assess the conditioning effects of phonological context,
grammatical function and speaker characteristics on the occurrence of these
weakening processes in a quantitative fashion. Their frequency in formal and
more casual speech styles is also compared.
As one of very few studies on Salvadorian Spanish, the first examination of the
Spanish spoken in a Latin American immigrant community in Canada and as an
investigation of two of the most researched phenomena in Spanish phonetics, this
volume will be of interest to specialists in Hispanic Linguistics and Spanish
Dialectology. It will also appeal to a much wider audience of phonologists and
sociolinguists as a study of consonantal weakening that contributes to our
general understanding of the influence of phonetic conditioning of lenition,
tests the hypothesis that sounds with morphosyntactic functions tend to resist
deletion and explores the significance of non-standard pronunciations in a
context characterized by dialect and language contact.
The first two chapters introduce the study. Chapter 1 provides an overview of
phonetic and phonological variation in Latin American Spanish and situates
Salvadorian Spanish within the context of other dialects. Major works on the
Spanish spoken in the United States are discussed, including studies of
characteristics that may reflect English influence and recent work on the
potential for dialect leveling and koineization resulting from contact between
multiple varieties of Spanish in immigrant communities. The author describes
Latin American immigration to Canada in general and Toronto’s Salvadorian
population in particular, noting the lack of previous studies on the Spanish
spoken in this city where, just as in urban areas of the U.S., multiple dialects
of the language are in contact with one another and with English.
In the second chapter, Hoffman describes her subjects and discusses her data
collection techniques in detail. A total of 30 informants, 14 men and 16 women
recruited at a community center for Spanish speakers in Toronto and at local
high schools, participated in traditional sociolinguistic interviews including
guided conversation as well as the reading of a short passage, a sentence list
and a word list. All were between 15 and 25 years of age. Although subjects'
socioeconomic circumstances in Toronto were very similar due to their immigrant
status, their families had represented a range of backgrounds and educational
attainment when they lived in El Salvador. These indications of pre-emigration
social position served to divide participants into status-based groups. Subjects
also varied with respect to their age at the time of their arrival in Canada as
well as the duration of their stay in the country. These differences were
utilized as a means of assessing the potential effects of dialect contact on
realizations of /s/ and /n/.
Chapter 3 presents patterns of /s/ variation in conversational speech and
describes the effects of phonetic factors, grammatical function, word frequency
and speaker characteristics on aspiration and deletion. It was observed that 49%
of syllable-final and word-final /s/s were retained while 28% were aspirated and
23% deleted. Weakening patterns were found to differ according to word
position. Sixty-seven percent of word-internal syllable-final /s/s were produced
as [s] while 28% of these tokens were aspirated and only 6% were deleted.
Word-finally, 46% of /s/s were realized as [s], 28% as [h] and 26% as [Ø].
Several phonetic contexts were found to have significant effects on /s/
weakening. There was tendency for preceding high vowels, following coronal
consonants and pauses to favor /s/ retention while following non-coronal
consonants promoted aspiration and voiced consonants, coronals and continuants
were associated with /s/ deletion. Both aspiration and deletion were more likely
to occur in unstressed syllables. Following unstressed syllables favored
deletion but had no effect on aspiration. The author notes that many of these
results appear to be coarticulatory effects. /s/ is retained when adjacent
sounds require tongue tip raising in the general area of the aveolar ridge but
tends to be glottalized or elided in other contexts. She attributes /s/ deletion
before coronal consonants to the Obligatory Contour Principle and explains the
frequent occurrence of [Ø] before voiced consonants as a result of both
articulatory and acoustic factors, noting that Spanish /s/ is realized as [z], a
voiced sibilant that is both more difficult to produce and also to perceive than
[s], before voiced consonants.
As Salvadorian Spanish is one of the few Spanish varieties known to exhibit
syllable-initial /s/ aspiration, the contextual factors associated with this
pronunciation were also examined. A variable rule analysis indicated that
preceding non-high vowels and the location of /s/ in unstressed syllables
promoted aspiration while preceding high vowels and pauses were associated with
[s]. These findings are similar to the only other quantitative investigations
of syllable-initial /s/ aspiration in Spanish that have been conducted to date,
Brown and Torres Cacoullos’ work (2001) on the speech of Chihuahua, Mexico and
Brown’s (2005) study of New Mexican Spanish.
The author also relates her findings regarding the effects of phonetic and
phonological conditioning of /s/ reduction to proposals about the phenomenon’s
original locus and subsequent spread to other contexts. Lipski (1985) has
suggested that, in Central America, /s/ weakening may have begun in word final
pre-consonantal position and spread, first to word-internal pre-consonantal
contexts, then to pre-pausal position, finally affecting pre-vocalic word /s/.
Hoffman did find higher rates of deletion in word-final pre-pausal versus
word-final pre-vocalic position. However, rates of word-final and word-internal
aspiration were nearly equal and, word-finally, aspiration was more frequent
before a following vowel than before a pause. Her results thus offer only mixed
support for this hypothesized contextual progression. It has also been argued
(Lipski 1999, Penny 2000) that syllable-initial aspiration results from the
spread of /s/ weakening from syllable-final pre-vocalic position to
syllable-initial pre-vocalic contexts. Hoffman observes that there are parallels
between the linguistic factors conditioning syllable-final and syllable-initial
pre-vocalic aspiration and finds that the individual speakers who most
frequently produced [h] in syllable-final pre-vocalic position also aspirated
syllable-initial pre-vocalic /s/ most often. Both results are compatible with
the claim that syllable final weakening is a precursor to syllable initial
As the Spanish the plural suffix and some of the language’s verbal inflexions
contain word-final /s/, Hoffman explored the relationship between deletion and
morphological status, testing the functional hypothesis that linguistic material
whose absence is likely to create ambiguity and potentially disrupt
communication resists deletion. She evaluated the effect of functional
constraints on /s/ elision in several ways, including comparisons of deletion
rates for monomorphemic words and inflections and an examination of the
relationship between the presence of disambiguating information and /s/
omission. Deletion rates for the two allomorphs of the Spanish plural, /s/ and
/es/, were also tabulated since, from a functionalist perspective, /s/ elision
should be more likely to affect the /es/ form in which the absence of /s/ would
not leave the plural completely unmarked. In addition, Hoffman noted the
strength with which phonological factors conditioned word-final /s/ deletion in
monomorphemes and in inflections as measured by differences between factor
weights within factor groups. She assumed that, if functional effects also
influence deletion rates for inflections, phonological factors should have less
impact on /s/ elision in plurals and verbal morphology. Her findings failed to
support a functionalist explanation of /s/ deletion as /s/ was actually more
often elided when its absence was most likely to obscure meaning. For example,
/s/ was slightly more likely to be deleted when part of inflectional morphology
than when in word-final position in monomorphemes and plural /s/ was more likely
to be deleted in noun phrases when preceding instances of plural /s/ had also
been elided. Hoffman notes that the latter result is similar to findings
reported by Poplack (1986) who referred to the phenomenon as ''string level
concord''. Furthermore, while the results of variable rule analysis did indicate
that phonological factor groups exerted more influence on deletion in
monomorphemes than in plurals, elision in verbal inflexion was greatly affected
by phonological factors.
Hoffman also examined patterns of deletion in very frequent words that exhibit
high rates of /s/ elision and were not included in the corpus employed to
investigate other aspects of /s/ weakening. She found that phonological factors
had a weaker influence on /s/ deletion in these items, including the very common
words 'pues' (well, then), 'digamos' (let's say) and 'entonces' (then), which
she terms “lexical exceptions” than on members of other lexical categories.
Hoffman interprets this result as support for Guy's (2007) proposal that
frequent words exhibiting high levels of phonetic reduction have two alternative
underlying representations, one intact and one lenited. Under this view, the
phonological context has no effect on deletion processes when speakers select
the reduced form of a word because this entry in the mental lexicon already
contains a /Ø/.
In the final section of Chapter 3, Hoffman describes the effects of social
factors on /s/ weakening and the possible social significance of /s/ aspiration
and deletion in the Salvadorian immigrant population of Toronto. All evidence
suggests that [s] is the prestige variant in this community, which is not
surprising given that [s] is also the standard pronunciation. Women and those
whose parents had attained higher levels of education preferred the unreduced
[s] in conversational speech while men and those who parents had completed fewer
years of formal study used [h] and [Ø] more often. Furthermore, speakers were
consciously aware of /s/ weakening and reported that they associated it with
rural origin and poor education. Hoffman also observes that [s] was used most
often by speakers who arrived in Toronto at earlier ages as well as by those who
had spent more time in Canada and speculates that this trend could be the result
of contact with /s/ retaining varieties of Spanish. Finally, she notes that the
two subjects who appeared to most strongly identify with their Salvadorian roots
exhibited high rates of [Ø] while two who commented on the fact that people did
not normally realize they were Salvadorian used [s] almost exclusively. Based on
this contrast, Hoffman conjectures that /s/ weakening may to some degree
function as a symbol of Salvadorian identity in the Toronto immigrant community.
Chapter 4 presents the effects of phonological, grammatical and social factors
on realizations of word final /n/. Variants considered included [n], [ŋ], and
[Ø] with or without nasalization of the preceding vowel. The [ŋ] accounted for
31% of all tokens while deletion with and without nasalization of the preceding
vowel was very infrequent, affecting only 10% of /n/s in the corpus. Hoffman
reports that preceding front vowels favored the [ŋ] and back vowels the [n], an
effect she attributes to the Obligatory Contour Principle. Preceding high vowels
disfavored elision, presumably because the elevated tongue position involved in
their production supports the articulatory movments needed for both /n/ and /ŋ/.
Following velar consonants, vowels and pauses favored [ŋ]. [Ø] was most likely
to occur before a pause or a non-continuant consonant. There was a
non-significant tendency for more velarization to occur in unstressed syllables,
but deletion was not affected by stress.
Since word final /n/ forms part of many Spanish verbal suffixes, the influence
of the sound's grammatical role on deletion was explored as a further test of
the hypothesis that phonemes with a morphological function resist elision. As in
the case of /s/, this proposal was assessed by several types of analyses,
including comparisons of deletion rates in monomorphemes and verbal suffixes,
examination of deletion in verb tenses for which /n/ elision would create three
way versus two way ambiguity and calculation of the strength with which
phonological factors conditioned deletion in verbs and monomorphemes. Results
for /n/ did not indicate that morphological function inhibits the deletion of
phonological material. However, as the author observes, the low rate of /n/
deletion in the sample makes it difficult to draw definite conclusions regarding
the relationship between this weakening process and morphological function.
Finally, the social conditioning of /n/ velarization and elision is explored.
Findings suggest that, unlike /s/ reduction, /n/ weakening may operate below the
level of speakers' conscious awareness. Hoffman reports that, when questioned
about the characteristics of Salvadorian Spanish, none of her informants
mentioned the pronunciation of /n/ as a distinctive aspect of their dialect. As
she observed that the velar variant was used slightly more often by women and by
those whose parents had completed more years of formal schooling, she concludes
that this pronunciation does not have negative associations for the Salvadorian
community in Toronto. Final nasal deletion, on the other hand, was used more
frequently by speakers with less educated parents. Noting that those who had
spent more time living in Toronto were more likely to use [ŋ], Hoffman
hypothesizes that this pronunciation might be symbolic of ''pan-Hispanic'' identity.
Chapters 5 and 6 primarily recapitulate and synthesize results already
presented, although Chapter 5 does describe the differences between realizations
of /s/ and /n/ across speech styles. Findings from this measure of relative
prestige are consistent with those obtained from analyses of conversational
speech; /s/ exhibited a great deal of style shifting while /n/ did not. Rates of
/s/ retention were much higher in more formal styles and reached one hundred
percent in the word list reading task. In contrast, the realization of final
nasals was very similar in all contexts with the exception of word list reading
which produced a high rate of [Ø]. Hoffman points out that, while the frequent
use of [Ø] in word reading probably reflects the tendency for deletion to occur
before a pause, this result supports her contention that /n/ reduction has
little social meaning in the Toronto Salvadorian community.
This volume makes several contributions to Hispanic Linguistics in particular
and linguistics in general. Firstly, as the conditioning of /s/ and /n/
weakening has been shown to vary across dialects (e.g. Brown 2008, Lipski 1986),
this detailed description of the processes in Salvadorian Spanish is a useful
addition to our knowledge of the phonetic patterns manifested in Spanish
varieties. Secondly, as findings are used to test hypotheses regarding the
progressive spread of weakening processes through different contexts and about
the possible functional conditioning of sound deletion, the study also
contributes to our understanding of the development and potential motivations
for sound change. Thirdly, the data presented regarding the social conditioning
of /s/ and /n/ weakening the Toronto Salvadorian community provide valuable
information about the possible effects of dialect contact on the social
significance of non-standard speech characteristics. The relationship between
social factors and consonantal variation revealed by Hoffman's investigation are
complemented by the findings of two more recent examinations of /s/ weakening
(Aaron & Hernández 2007) and /n/ reduction (Hernández 2009) in the Salvadorian
population of Houston, Texas. These two groups exhibit somewhat different
patterns of /s/ and /n/ variation, apparently as a response to differences in
the social dynamics and dialectal characteristics of their Hispanic communities.
Comparisons between them thus provide an opportunity to examine the specific
mechanisms that determine the outcomes of dialect contact. Finally, the author's
thorough examination of contextual factors, including several not investigated
in many studies of /s/ and /n/ weakening, and her use of variable rule analysis
to provide quantitative results are both very positive aspects of this study.
This is a welcome addition to Hispanic Linguistics and variationist
sociolinguistics, and I have only one minor criticism: While Hoffman’s
explanation of the relatively weak effects of phonological factors on /s/
weakening in very frequent discourse markers in terms of Guy’s (2007) hypothesis
does account for this result, it is surprising that no mention was made of the
alternative explanation of this pattern offered by usage based phonology (Bybee
2001, Pierrehumbert 2003) and applied to similar results in recent dissertations
by Minnick Fox (2006) and Brown (2008). Both perspectives probably fit her data
equally well, and it would have been worthwhile to at least mention if not
explore both possibilities.
Aaron, Jessie Elana and Hernández, José Esteban. 2007. Quantitative Evidence for
Induced Accommodation: Shifts in /s/ Reduction Patterns in Salvadoran Spanish in
Houston. Spanish in Contact: Policy, Social and Linguistic Inquiries, ed. by K.
Potowski and R. Cameron, 329-344. Amsterdam: John Benjamins.
Brown, Earl Kjar. 2008. A Usage Based Account of Syllable- and Word-final /s/
Four Dialects of Spanish. Ph.D. dissertation: University of New Mexico.
Brown, Esther. 2005. New Mexican Spanish: Insight into the Variable Reduction of
inihial” (-s). Hispania 88(4), 813-824.
Brown, Esther and Torres Cacoullos, Rena. 2001. Spanish /s/: A Different Story
from Beginning (initial) to End (final). A Romance Perspective on Linguistic
Knowledge and Use: Selected Papers from the 31st Linguistic Symposium on Romance
Languages, ed. by R. Nuñez Cedeño, L. López, and R. Cameron. Amsterdam: John
Bybee, Joan. 2001. Usage Based Phonology. Cambridge: Cambridge University Press:
Guy, Gregory R. 2007. Variation and Phonological Theory. Sociolinguistic Variation:
Theories, Methods, and Applications, ed. by R. Bayley and C. Lucas, 5-23. New York:
Cambridge University Press.
Hernández, José Esteban. 2009. Measuring Rates of Word-Final Nasal Velarization:
of Dialect Contact on In-Group and Out-group Exchanges. Journal of
Sociolinguistics, 13(5), 582-612.
Lipski, John M. 1985. /s/ in Central American Spanish. Hispania, 68, 143-9.
Lipski, John M. 1986. Reduction of Spanish Word-final /s/ and /n/. The Canadian
Linguistics, 31(2), 139-156.
Lipski, John M. 1999. The Many Faces of Spanish /s/-Weakening: (Re)alignment and
Ambisyllabicity. Advances in Hispanic Linguistics, ed. by J. Gutiérrez Rexach
and F. Martínez Gil, 198-213. Sommerville, MA: Cascadilla Press.
Minnick Fox, Michelle Annette. 2006. Usage-Based Effects in Latin American Spanish
Syllable-final /s/ Lenition. Ph.D. dissertation, University of Pennsylvania.
Penny, Ralph. 2000. Variation and Change in Spanish. Cambridge: Cambridge
Pierrehumbert, Janet. 2003. Phonetic Diversity, Statistical Learning and
Phonology. Language and Speech 46.2-3, 115-154.
Poplack, Shana. 1986. Acondicionamiento gramatical de la variación fonológica en
puertorriqueño. Estudios sobre la fonología del español del Caribe, ed. by R.
Nuñez Cedeño, I. Páez, & J. Guitart, 97-107. Caracas: La Casa de Bello.
| ABOUT THE REVIEWER:
Ann Marie Delforge is Assistant Professor of Hispanic Linguistics at
Montclair State University. She is interested in sociophonetics,
phonological theory and language contact phenomena.