The Effect of Dialect Familiarity via a Study Abroad Experience on ...

The Effect of Dialect Familiarity via a Study Abroad
Experience on L2 Comprehension of Spanish

Lauren Beth Schmidt
Indiana University


1. Introduction

Speaker dialect has been found to play an important role in listening comprehension by learners of
a second language (L2) (see, for example, Brown, 1968; Eisenstein & Berkowitz, 1981; Major et al.,
2005; Wilcox, 1978). L2 listeners have been tested on their comprehension of various international,
regional, and nonnative varieties of the target language, and differences in comprehension have been
found according to dialect. It has been proposed by Tauroza and Luk (1997) that these differences in
target language comprehension according to dialect are in fact due to dialect familiarity, with greater
comprehension of those varieties to which the learner is exposed.
This study addresses the role of the target language variety, or dialect, on listening comprehension
of Spanish as a second language. Specifically, the goal of the study is to determine if exposure to a
previously unfamiliar dialect of Spanish through a three-week study abroad program in the Dominican
Republic has an effect on the performance on listening comprehension tasks by L2 learners of Spanish.
Previous studies on the effect of dialect on listening comprehension by L2 learners have
frequently considered English as the target language, but no previous study has examined the effects of
dialect on listening comprehension of Spanish as the L2. Also, this study is the first to offer an
examination of the effects of exposing L2 learners to a previously unfamiliar variety of the target
language by following the learners’ listening comprehension abilities from pre-exposure to post-
exposure to the unfamiliar variety via a study abroad program in the targeted dialectal region.
The following section presents an overview of previous work on the role of study abroad in L2
listening comprehension as well as the relationship between target language dialect and L2 listening
comprehension. Section 3 describes the phonological characteristics of the targeted Spanish variety
spoken in Santo Domingo, the Dominican Republic. Section 4 presents the methodology employed in
the current study, and Section 5 describes the results found. A discussion of the results and final
conclusions are presented in Sections 6 and 7.

2. Listening comprehension
2.1. Effects of a study abroad experience on L2 listening comprehension

Despite the range of studies investigating the effect of a study abroad experience on the
acquisition of such areas of second language acquisition such as morphosyntax, pragamatics, and oral
fluency (see Lafford & Collentine, 2006, for an overview of study abroad research for Spanish), there
have been very few studies of the effects of study abroad on listening comprehension of the L2. Allen
and Herron (2003) found that after a study abroad experience, learners of French had a higher level of
confidence in their ability to perform certain listening tasks and significantly improved in a multiple-
choice task that tested their comprehension of three scenes from a French police TV drama. Huebner
(1995) determined that beginning learners of Japanese who participated in a study abroad experience
in Japan tested better in listening and reading comprehension than those beginning learners of Japanese
who took classes at the home institution. Huebner claimed that the better performance by the study
abroad group was due to more opportunities to practice the listening and reading skills.
*I would like to thank Kimberly Geeslin for her guidance throughout all stages of the research and to thank Erik
Willis for the use of his database to construct the tasks. I would also like to thank them both as well as two
anonymous readers for their detailed and insightful comments and suggestions on an earlier version of the paper.
© 2009 Lauren Beth Schmidt. Selected Proceedings of the 11th Hispanic Linguistics Symposium, ed. Joseph
Collentine et al., 143-154. Somerville, MA: Cascadilla Proceedings Project.

---

144
In summary, the effects of a study abroad experience on L2 listening comprehension have not
been investigated in detail, and only general findings regarding overall comprehension have been
determined. This study adds further insight into the relationship between an abroad experience and
listening comprehension as well as addresses several factors that may come into play to explain effects
of a study abroad program.

2.2. Dialect familiarity and L2 listening comprehension


There is much empirical support that speaker dialect has an effect on listening comprehension,
even for L2 listeners. For example, Major et al. (2005) determined that English as a second language
(ESL) listeners scored lower on listening comprehension tests when hearing ethnic and international
dialects of English than when hearing standard American English. Eisenstein & Berkowitz (1981)
found that standard English was more intelligible for advanced beginners and high intermediate-level
learners than working-class New Yorkese or foreign-accented English. However, it is not always the
standard dialect that L2 listeners are most successful in comprehending. Chinese learners of English in
China were most successful in comprehending Singaporean-Malaysian English than British English
Received Pronunciation (RP), General American English, or standard Australian English (Wilcox,
1978). Twi and Ewe speakers in Africa best understood Twi-accented or Ewe-accented English,
respectively, than British English RP or the English spoken by the other tribe (Brown, 1968).
Tauroza and Luk (1997) claim that the fact that some learner populations are more successful in
comprehending a standard variety while other learner populations achieve greater comprehension of a
L1-accented variety is explained by the influence of dialect familiarity. They propose that L2 learners
understand best the dialect(s) to which they are exposed, be it the standard variety as in the case of
learners in the target language environment (Eisenstein & Berkowitz, 1981; Major et al., 2005) or be it
a variety spoken by speakers of the same first language, as may be the case of learners in a foreign
language environment (Brown, 1968; Wilcox, 1978).
Interestingly, however, Fox and McGory (2007), in a study of listening comprehension of
Standard American English and Southern American English by L2 listeners, native speakers of
Japanese, who were residing in one or the other dialectal regions, found that exposure to the Southern
American English variety did not improve the learners’ perception of southern vowels. In fact, both
groups of learners identified the Southern vowels significantly worse than the standard vowels.
However, Fox and McGory suggest that there may be sociolinguistic factors contributing to this
unexpected result such as the variety spoken by the English instructors, attitudes toward nonstandard
dialects of the target language, and the amount of time that the learners actually interact with
individuals who speak the southern variety.

3. Dominican Spanish


Dominican Spanish shares several features with that of its surrounding Caribbean and coastal
neighbors. Several syllable-final consonants undergo weakening and deletion processes, which may
cause additional difficulties for L2 learners in comprehension of the Dominican speech. These include
the aspiration or, more commonly, the complete deletion of the syllable-final and word-final sibilant
/s/; the deletion of the intervocalic and word-final dental approximant [δ]; and the variation of the
syllable-final liquids /ɾ/ and /l/ (Alba, 2004; Jiménez Sabater, 1975; Lipski, 1994). Other phonological
characteristics of the Dominican dialect include the velarization of word-final nasals1, a weak
aspiration of the fricative /x/ (Alba, 2004; Jiménez Sabater, 1975), and pre-breathy voicing of the trill
(Willis, 2006; 2007a). A number of morphological, syntactic, and lexical features also distinguish the
Dominican dialect (see Lipski, 1994). However, these features were not included in the current study
and were controlled for in the speech samples included2.

1 Nasalization of the preceding vowel and deletion of the nasal consonant has also been attested in Dominican
Spanish (Hualde, 2005; Jiménez Sabater, 1975).
2 Dominican prosody has also been found to differ from that of other dialects of Spanish (Willis, 2007b). It is
recognized that Dominican dialectal prosodic patterns may be present in the samples included at the phrase-level.

---

145

Finally, within the Dominican Republic, three or four different dialectal zones can be proposed:
the northern Cibao region; the region surrounding Santo Domingo, the capital; the extreme eastern tip
of the island; and various Afro-Hispanic enclaves (Lipski, 1994). What distinguishes the dialectal
zones is the realization of syllable-final liquids. The dialect to which the study abroad learners were
most exposed was the Santo Domingo speech, which is characterized by lambdacism, a lateral
realization [l] of the rhotic /ɾ/, as in the word for woman: mujer [mu.hél] (Alba, 2004).

4. Methods
4.1. Research questions

The following are the research questions posited in the present study:
(1) How does exposure to a previously unfamiliar dialect of the L2 through a study abroad
experience affect L2 listening comprehension of that dialect at the word- and phrase-levels?
(2) How does discourse level (word-level versus phrase-level) play a role in L2 listeners’
comprehension of the Dominican speech?
(3) What role does each Dominican phonological feature play in the L2 learners’ listening
comprehension of Dominican speech?

4.2. Participants


Eleven participants were included in this study from an original participant pool of eighteen. Two
of the study abroad participants were eliminated at the beginning of the study as they were novice or
low beginners and did not have sufficient competence in the L2 to complete the listening tasks. Five
additional participants were later eliminated from the analysis for failure to complete both tests.

The participants spanned a range of proficiency levels in Spanish, from low intermediate through
near-native; however, the majority classified themselves as intermediate or advanced. There was one
bilingual participant who was raised in a Mexican Spanish-speaking household in the United States
(participant 8). Another participant (participant 2) had a parent from the Dominican Republic, but she
did not consider herself bilingual and the quantity of Spanish heard and spoken in the household is
uncertain. Two of the participants had also previously spent over a year abroad in other dialectal
regions of the Spanish-speaking world3. All participants were native speakers of English and had
limited or no knowledge of any other languages (other than Spanish).

4.3. The study abroad program

The program consisted of a week of readings and class discussion at the home institution prior to
leaving for the Dominican Republic in preparation for the program. Three weeks were then spent in
the capital city, Santo Domingo, where the students stayed together in a hotel. The focus of the
program was on Dominican culture and literature, yet as the students came from a diverse background
of proficiency levels of Spanish, the majority of the readings and class discussions were in English.
However, the study abroad participants were exposed on a daily basis to Dominican Spanish through
talks by invited guest lecturers and tour guides, instruction by the program coordinator, exposure to
Dominican television and radio, and daily interactions with Dominicans in public places such as
restaurants, shops, the market, the hotel, and so on. Although the program did not include any
instruction on the Spanish language, prior to leaving for the Dominican Republic (and after completing
the pretest), an invited expert on Dominican Spanish gave a lecture and played examples of Dominican
dialectal features, including Dominican phonology, lexicon, and syntax.

4.4. Tasks
4.4.1. Dialectal listening comprehension tasks


In order to address the research questions posited above, a series of listening comprehension tasks
were completed by the study abroad participants one week before leaving for the Dominican Republic

3 Participant 1 spent a year in Burgos, Spain, and four months in Salamanca, Spain, and participant 7 studied one
year in Argentina.

---

146
and then again two days after returning from the three-week time abroad. The series of tasks consisted
of a word comprehension task, a phrase comprehension task, and a discourse comprehension task.
Only the results from the word and phrase comprehension tasks are included in the present study.

The participants completed the series of listening comprehension tasks at individual computer
stations in a computer language laboratory. After hearing a spoken word or spoken phrase twice by
clicking on the corresponding number of the speech sample in a PowerPoint presentation, the
participants wrote down the word or phrase that they heard in Spanish and then wrote the word’s or the
phrase’s meaning in English. The word task was completed first and consisted of twenty-five spoken
words in Spanish. The phrase task consisted of fourteen spoken phrases in Spanish.
The same speech samples were included in both the pretest and the posttest, but presented in a
different order for each section. The same samples were used in order to ensure that each test was of
equal difficulty. The assumption was made that the rearranged speech samples and the four-week
interval between the two tests would aid in diminishing potential effects of the repeated test.

4.4.1.1. Word task


Nineteen of the twenty-five words presented in the Word Task were extracted from sociolinguistic
interviews with five Dominican male speakers between the ages of twenty to thirty-five and of low to
low-middle class (Willis, 2001). The remaining six words came from interviews with three university-
educated male native speakers of Spanish of the same age from other regions of the Spanish-speaking
world: from Bogota, Colombia, from San Isidro de Pérez Zeledon, Costa Rica, and from Seville,
Spain. For the purposes of this study, we will term the samples extracted from these three speakers as
“standard Spanish”, in the sense that the phonetics of the specific samples chosen do not reflect
dialectal variants4.

Of the nineteen words extracted from the interviews with the Dominican speakers, originally eight
words displayed syllable- or word-final /s/ deletion, four words lambdacism, four words word-final
nasal velarization, and two words word-final /d/ deletion (see Appendix A). However, three of the
words with word-final /s/ deletion were later excluded from the analysis when the final /s/ was the
plural morpheme, as without surrounding syntactic, semantic, and discourse clues, it was not possible
for the listeners to capture the notion of plurality.

The six words spoken by the standard speakers were chosen as these six samples did not reflect
any dialectal phonological features (see Appendix A). Thus, the assumption was made that no features
of the standard speech would be unfamiliar to the participants.

4.4.1.2. Phrase task


Of the fourteen phrases presented to the learners, ten were taken from the interviews with the
Dominican speakers and the remaining four from interviews with the three other Spanish speakers5
(see Appendix B). Of the ten phrases spoken by Dominican speakers, there were four examples of
phrases with one or more words exhibiting /s/ deletion, three with lambdacism, two with /d/ deletion,
and one phrase with final nasal velarization. The two phrases that contained words with the deleted
plural morpheme /-s/ were included in the analysis at the phrase-level as the notion of plurality can be
determined by the surrounding semantic, syntactic, or discourse information found at the phrase level.

4 It is recognized here that these dialects of Spanish, particularly the Seville and San Isidro variants, are not
typically considered to represent ‘standard’ varieties of Spanish, and in fact, as pointed out by a reviewer,
Andalusian Spanish shares many features with Caribbean varieties. However, here they are referred to as
representing a ‘standard’ comparison due to the careful selection of speech samples that do not exhibit dialectal
variation. The current study, however, makes no claims as to the ‘standardness’ of these dialects as a whole.
5 Care was taken to insure that no dialectal phonological features were present in the phrases taken from the three
speakers representing other Spanish dialects. However, post hoc analysis revealed the presence of the retroflex r
in the one phrase spoken by the Costa Rican speaker.

---

147
4.4.2. Language background and proficiency measures

The day of the pretest the study abroad participants completed a language background
questionnaire in which they answered questions concerning their knowledge of other languages,
number of years learning Spanish and the last Spanish class taken, the type and amount of contact with
Spanish speakers outside of the classroom, and previous time abroad in a Spanish-speaking region.
The participants also self-ranked their proficiency in speaking, writing, listening, and reading Spanish
on a nine-point scale from novice to native-speaker. The four rankings were later averaged for each
participant to assign each participant an overall self-identified proficiency level in Spanish.

4.4.3. Vocabulary recognition task


The participants completed a vocabulary recognition task as the final task in the posttest. The
students were presented with forty words written in Spanish that had been presented on the word task,
the phrase task, and the discourse task and were asked to mark if they (a) knew the meaning of the
word, (b) had heard the word previously but did not know its meaning, or (c) had never heard or seen
the word before. If the participants marked that they knew the word, they were instructed to write
down its equivalent in English. The participant responses to the vocabulary recognition task were later
analyzed in order to determine the learners’ familiarity with the words presented in the speech samples
in the listening comprehension tasks.

5. Results
5.1. Listening comprehension: Word task


In order to address the first part of Research Question 1 regarding the effect of exposure to a
previously unfamiliar dialect of the L2 through a study abroad experience on comprehension of that
dialect at the word-level, the number of words spoken by the Dominican speakers that were
comprehended was divided by sixteen, the total number of words included in the analysis, for each
participant and for each test (pretest and posttest). A score was also given to each participant for both
tests according to the number of comprehended words of the standard speech divided by six, the total
number of words spoken by the standard speakers. The results of the Word Comprehension Task
according to participant, dialect, and test are presented in Table 1.

Table 1. Word Task Listening Comprehension Scores According to Dialect and Test
__________________________________________________________________

Dominican (16)



Standard (6)



_________________


________________
Participant Pretest

Posttest


Pretest
Posttest
__________________________________________________________________
1



69%

88%




100%

100%
2



44%

63%




67%

83%

3



25%

63%




67%

83%
4



38%

44%




67%

83%
5



69%

56%




100%

100%
6



56%

69%




83%

83%
7



44%

69%




67%

83%
8ª


88%

100%




100%

100%
9



44%

50%




83%

67%
10


31%

38%




100%

83%
11


19%

13%




17%

33%
__________________________________________________________________
Mean


47.9%
59.4%



77.4%
81.6%
__________________________________________________________________
(N) = number of words presented. ªThe bilingual speaker.

---

148
When we consider the results of the comprehension of the Dominican speech on the pretest and on
the posttest (after the study abroad experience) as displayed in Table 1, we see that the majority of the
participants (all but participants 5 and 11) were more successful in comprehending the Dominican
speech at the word-level after the experience abroad, reflected by the higher scores on the posttest6. A
paired-samples t-test was run using SPSS software to determine if the difference between the
comprehension scores of the Dominican speech on the pretest and the posttest was significant. Indeed,
it was found that the increase in comprehension scores of the Dominican speech at the word-level after
the study abroad experience was significant [t(10) = -2.695, p < .05]. We also note in Table 1 that the
participants appear to perform better on the comprehension of the standard Spanish speech on the
posttest than on the pretest. However, a significant difference was not found [t(10) = -1.11, p = .292].

Finally, it is important to note that the L2 listeners scored higher on the comprehension of the
standard speech (77% and 82% comprehension on the pretest and posttest, respectively) than of the
Dominican speech (48% and 59%, respectively) at the word-level. The listening comprehension scores
of the Dominican speech on the word task were found to be significantly lower than those of the
standard speakers for both the pretest [t(10) = -5.46, p < .05] and the posttest [t(10) = -5.16, p < .05].

5.2. Listening comprehension: Phrase task

The second part of Research Question 1 regarding the effect of exposure to a previously
unfamiliar dialect of the L2 through a study abroad experience on comprehension of that dialect at the
phrase-level was addressed by assigning each participant comprehension scores of the Dominican and
standard Spanish speech for both tests at the phrase-level. The comprehension scores were assigned by
dividing the number of phrases comprehended by ten (the total number of phrases extracted from the
Dominican interviews) or by four (the total number of standard Spanish phrases presented)7. The
results of the Phrase Task according to dialect, test, and participant are presented in Table 2.

Table 2. Phrase Task Listening Comprehension Scores According to Dialect and Test
__________________________________________________________________

Dominican (10)



Standard (4)



_________________


________________
Participant Pretest

Posttest


Pretest
Posttest
__________________________________________________________________
1



40%

70%




100%

100%
2



10%

20%




75%

75%
3



10%

30%




50%

75%
4



10%

10%




50%

25%
5



30%

40%




75%

75%
6



0%

40%




50%

75%
7



20%

30%




75%

75%
8ª


60%

70%




100%

100%
9



20%

30%




50%

25%
10


10%

20%




25%

75%
11


0%

0%




0%

0%
__________________________________________________________________
Mean


19%

33%




59%

64%
__________________________________________________________________
(N) = number of phrases presented. ªThe bilingual speaker.

6 There is a large amount of variation in comprehension scores of the Dominican speech at the word-level, which
may reflect the range of L2 proficiency levels of the participants. The results from a Pearson Correlation found a
significant strong positive correlation between proficiency level (intermediate, advanced, high advanced, (near)
native) and comprehension scores of the Dominican speech at the word-level [r(9) = .83, p =.002].
7 Participants were given the point for a comprehended phrase only if there was comprehension of the entire
phrase.

---

149

When we consider the results for the comprehension of the Dominican speech at the phrase-level,
we note in Table 2 that as with the results of the comprehension of the Dominican speech at the word-
level, the majority of the participants scored higher in comprehension of the Dominican speech at the
phrase-level on the posttest than on the pretest8. Only participant 4 remained the same and participant
11 did not comprehend any Dominican phrase on either of the tests. A paired samples t-test found the
difference in pretest and posttest scores for the comprehension of the Dominican speech at the phrase-
level to be significant [t(10) = -3.75, p < .05].

The group’s comprehension score as a whole of the standard speech at the phrase-level also
increased from the pretest to the posttest (from 59% to 64% accuracy). However, contrary to the
significant increase in comprehension scores of the Dominican speech on the Phrase Task, the
difference between tests for comprehension of the standard speech at the phrase-level was not
significant [t(10) = - .69, p = .506].

Finally, it is important to also note that when the scores on the Phrase Task according to dialect
are compared, we see that the L2 listeners scored higher on the comprehension of the standard Spanish
than on the Dominican speech, for both the pretest (59% comprehension of the standard speech
compared to 19% of the Dominican) and the posttest (64% compared to 33%). This difference in
comprehension of the phrases according to dialect was found to be significant for both the pretest
[t(10) = -6.88, p < .05] and the posttest [t(10) = -5.07, p < .05].

5.3. Listening comprehension according to discourse level

Research Question 2 questioned the effect of discourse level (word-level versus phrase-level) on
the L2 learners’ listening comprehension of the Dominican speech. Table 3 restates the comprehension
scores of the Dominican speech according to discourse level for the group.

Table 3. Comprehension Scores of the Dominican Speech According to Discourse Level
__________________________________________________________________
Test




Word Task



Phrase Task

__________________________________________________________________
Pretest




48%





19%



Posttest




59%




33%



__________________________________________________________________


As displayed in Table 3, the L2 listeners performed better on the comprehension of the Dominican
speech on the Word Task than on the Phrase Task. Even the bilingual speaker (participant 8), who
scored a 100% on the Dominican Word Task posttest did not score higher than 70% on the Dominican
Phrase Task, as reflected in the presentation of the individual results in Tables 1 and 2 above. The
difference in comprehension of the Dominican speech according to discourse level was found to be
significant both for the pretest [t(10) = 8.49, p = .000] and for the posttest [t(10) = 8.73, p = .000].

5.4. Analysis of factors contributing to difficulties in comprehension of Dominican speech
5.4.1. Effect of phonological feature present


The effect of the presence of specific phonological features on the comprehension of the
Dominican speech was considered (Research Question 3). To determine the effect of each Dominican
phonological feature on comprehension of the Dominican speech by the L2 listeners, a comprehension
score of the words spoken by the Dominican speakers was calculated for each participant according to
the feature present9. Thus, for example, for the feature final nasal velarization, the total number of

8 Once again we find much variation in participant scores of the Dominican speech. A Pearson Correlation
determined a significant positive correlation between proficiency level (intermediate, advanced, high advanced,
(near) native) and comprehension scores of the Dominican speech at the phrase-level [r(9) = .86, p = .001].
9 Analysis of the effect of Dominican feature on listening comprehension was carried out with the Word Task data
only. The Phrase Task data were not included due to the small number of phrases representing each feature and to
the fact that many of the phrases had more than one such feature.

---

150
words exhibiting nasal velarization that participant X comprehended was divided by the total number
of words presented in the Word Task that displayed velarization. Only words with only one feature
were included in the analysis, as it is not possible to discern which feature may impede comprehension
of a word if more than one feature is present. Thus, two words, martes and asustado, could not be
included in this part of the analysis as they both exhibited more than one Dominican feature. Also, two
additional words, iban and también, were excluded as posthoc analyses revealed a standard alveolar
articulation rather than a velar articulation of these two word-final nasals. Thus, the analysis was based
on the comprehension scores of twelve words exhibiting Dominican phonological features at two
times, as the pretest and posttest results were combined for this part of the analysis. The results of the
comprehension of the Dominican speech on the Word Task according to feature present are presented
graphically in Figure 1.
Figure 1. Bar Graph of Comprehension Scores of Dominican Speech at the Word-level According to
Feature Present.
(N) = total number of speech samples representing each feature (pre- and posttests combined).
As displayed in Figure 1, there was an effect of phonological feature on the comprehension of the
Dominican speech at the word-level by the L2 listener group. The L2 listeners were least successful in
comprehension of the Dominican speech when the syllable-final /s/ was deleted (mean comprehension
at 32%) or when the rhotic tap was realized as a lateral (mean comprehension at 46%). The presence of
word-final /d/ deletion or nasal velarization did not seem to impede L2 listening comprehension to
such an extent (with 71% and 68% comprehension, respectively).
A one-way ANOVA was run to determine if the differences found in comprehension of the
Dominican speech at the word-level according to phonological feature present were significant. A
significant effect for feature was found [F(3, 40) = 4.853, p = .006]. However, post hoc Tukey tests
with adjusted p-values revealed a significant difference only between the features /d/ deletion and /s/
deletion (p = .048) and a difference approaching significance between nasal velarization and /s/
deletion (p = .08).
5.4.2. Ef ect of word familiarity
The responses of the L2 learners to the Vocabulary Task were analyzed according to response
type (word known, word familiar, word never seen before) for each of the sixteen words spoken by the
Dominican speakers and the six words spoken by standard speakers that were presented in the word
comprehension task.
Of the 16 words spoken by Dominican speakers, the average number of words with which each
participant was familiar (either knew the meaning or had claimed to have heard or seen it previously)
was 15.2 (SD = 2.5), while the average number of words of which each participant actually new the
meaning was 13.2 (SD = 1.3). The average number of the 6 words spoken by the standard speakers

---

151
with which the learners claimed familiarity was 5.5 (SD = 0.5), while the average number of which
each participant knew the meaning was 4.4 (SD = 0.8). Thus, it is safe to claim that the L2 listeners
knew or were at least familiar with the vast majority of lexical items presented to them and that
difficulties in comprehension of the speech must then be due to features of the speech itself rather than
to the fact that the learners were unfamiliar with the words.

6. Discussion
6.1. Effect of dialect exposure on listening comprehension through a study abroad experience


The results of this study support the claim made by Tauroza and Luk (1997) that dialect
familiarity has a significant effect on listening comprehension by L2 learners. After exposure to a
previously unfamiliar dialect of Spanish, here, Dominican Spanish, the L2 learners in this study
significantly improved in comprehension of the Dominican speech, both at the word-level as well as at
the phrase-level. The three-week study abroad program afforded an opportunity for the L2 learners to
become exposed to a dialect of the L2 with which they had no previous experience.
Also, the participants made slight gains in comprehension of the standard Spanish speech at the
word- and phrase-levels. An improvement in general comprehension of the L2 is not surprising, as
Huebner (1995) found that the increased opportunities to practice listening skills that an abroad
experience affords enabled the L2 learners of Japanese who studied in Japan in his study to perform
better than those learners who stayed at the home institution. However, the gain in comprehension of
the standard speech was not found to be significant. This finding is important in our analysis as it
provides support that the gain in comprehension of the Dominican speech did not simply reflect an
overall gain due to exposure to Spanish in general, but that it was due to the specific exposure to the
Dominican dialect.

It is also important to note that the L2 learners in this study were not only exposed to the Santo
Domingo variety of Spanish through the input heard during the three-week stay in the Dominican
Republic, but also through the explicit presentation of various phonological, morphosyntactic, and
lexical features of Dominican speech presented by a guest lecturer prior to the time abroad. It is
unknown to what extent it was the input heard by the learners or the explicit awareness of
phonological features of the dialect (or the combination of both) that contributed to a significant gain
in comprehension of Dominican speech. Future study should separate these factors (input, instruction,
and the combination of both input and instruction) to determine which may have the greatest effect on
dialect-specific listening comprehension.

6.2. Comprehension according to task type


The participants performed significantly better in comprehension of the Dominican speech on the
Word Task than on the Phrase Task. This could be due to the fact that five of the ten phrases had not
only one, but two Dominican features, which may have made comprehension of the Dominican speech
even more difficult for the L2 listeners. Also, seven of the ten phrases displayed /s/ deletion or
lambdacism, which were the two features found to most impede comprehension of the Dominican
speech by the L2 listeners. Finally, comprehension of the Dominican speech at the phrase-level may
also have been more difficult for the learners due to dialectal differences in prosody.
However, the L2 listeners also performed better on the comprehension of the standard Spanish on
the Word Task than on the Phrase Task, which would suggest alternatively that lower comprehension
of both the Dominican and standard speech on the phrase tasks is due to the length of speech heard (a
phrase as opposed to a single word). However, when the results for the comprehension of the standard
Spanish speech at the phrase level are examined, we see that many of the participants had difficulties
understanding one phrase in particular spoken by the Costa Rican speaker, mi tía es enfermera (my
aunt is a nurse)10, which, as there were only four test phrases included to begin with, may have been
the cause of such low scores on the comprehension of the standard Spanish at the phrase level. Had

10 The difficulty of the learners with this phrase may be due to the dialectal phonetic realization of the Costa Rican
tap /ɾ/ with a retroflex pronunciation, similar to the English rhotic (see Lipski 1994). This was the only case of a
dialectal feature in what are considered the standard speech samples in this study (see footnote 6).

---

152
there been more samples of standard Spanish phrases, we might expect to see higher scores for
comprehension of the standard speech on the Phrase Tasks.

6.3. Location of the phonological feature


Could it be the case that some features of the Dominican variety of Spanish cause more
difficulties for the L2 listeners than others? Results from this study suggest that different phonological
features of the Dominican Spanish have different effects on comprehension of the Dominican speech.
Deletion of the syllable-final sibilant /s/ and realization of the syllable-final tap /ɾ/ as the lateral liquid
[l] impede comprehension by the L2 listeners to a greater extent than do deletion of word-final /d/ or
velarization of the word-final nasal. It is proposed that this difference in effect of feature may be due to
the previous linguistic experience of the learners with Spanish phonetics and phonology in general.
First, the processes of /s/ deletion (as in “turiØta”) and lambdacism (as in “po[l]que”) generate
realizations of sounds that are not allophones of the underlying /s/ or /ɾ/ in either standard Spanish or
in the L1 (English), but, in fact, are associated with a different phoneme altogether. Thus, the L2
listeners may be unable to associate these Dominican phonetic realizations, Ø and [l], with their
corresponding phonemes, /s/ and /ɾ/. The learners must learn to re-associate these new allophonic
segments (or complete deletion of the segment) with the corresponding Spanish phonemes in the
appropriate contexts.
It is proposed that the phonological process of /d/ deletion, although similar to /s/ deletion in that
it involves the complete absence of a Spanish phoneme, presents a different case. In the majority of
Spanish dialects, the voiced stops /b d g/ are categorically “weakened” from a stop to a spirantized
approximant [β δ γ] in intervocalic position (Hualde, 2005). Thus, although in this case there is further
weakening of the spirantized approximant [δ] to complete deletion, the L2 listeners are already
familiar with a weakening process of the voiced stop /d/, possibly lessening the effect of a further
weakening of this segment (to complete deletion) on the comprehension of speech with this feature.
Finally, it is proposed that the velarization of the word-final nasal also causes less difficulty in
comprehension by the L2 listeners as the velar nasal [ŋ] never occurs as a phoneme in any dialect of
Spanish and categorically appears as an allophone in standard Spanish when preceding a velar
consonant, both word-internally and across boundaries (Hualde, 2005; Quilis, 1993). Thus, L2
listeners’ previous familiarity with both /d/ weakening and nasal place assimilation to following velar
consonants in other varieties of Spanish may explain why the deleted /d/ and velarized nasal caused
fewer difficulties in comprehension of the Dominican speech.

However, there is another possible explanation for the differences found in the effect of feature
present. Due to the nature of the speech samples included in the present study, there was an
unanticipated relation between specific feature present and location of the targeted feature, with all the
samples included in the Word Task of /d/ deletion and /n/ velarization occurring word-finally and all
the samples included in the Word Task of /s/ deletion and lambdacism occurring word-internally.
Thus, it is possible that those processes that occur within the word have a greater effect on L2 listener
comprehension than those that occur at the end of the word. In fact, Tauroza (1993) found that L2
learners of English were not affected in their comprehension of English when word-final consonants
were deleted in the speech of native speakers. Future study must separate these two factors in order to
determine if the effect on comprehension is due to the specific feature present, to the location of the
feature within the word, or to both.

7. Conclusion

The intermediate through near-native L2 learners of Spanish in this study significantly improved
in comprehension of the Dominican speech after the three-week study abroad experience in the
Dominican Republic, providing support to Tauroza and Luk’s claim that dialect familiarity has an
effect on listening comprehension by L2 listeners (1997). However, even after three weeks of intense
exposure to the Dominican dialect through a study abroad experience, the L2 learners in this study did
not reach the same level of accuracy of listening comprehension of the Dominican speech at either the
word-level or the phrase-level as of the more standard and previously familiar speech. However, we do

---