Adult acquisition of Spanish intonation: the effect of romantic relationships.
This paper uses the AM approach to analyze the phonetic characteristics of key intonational events in nonnative Spanish. The participants were five native English speakers involved in romantic relationships with native Spanish speakers. The results indicate that, to a certain degree, normative acquisition of Spanish intonation does occur. The speakers' successful acquisition can be attributed to their identification of their partners' dialect as the specific intonation target and their motivation to use Spanish as the primary' language of communication in the relationship. The results lend support to the importance of apperceived input in the process of second language acquisition (Gass 1988).
Keywords: intonation, autosegmental-metrical approach, Spanish, broad focus declaratives, second language acquisition
1. INTRODUCTION. Within the field of second language acquisition, little work has been done on the acquisition of Spanish intonation by adult native speakers of English. Nibert (2005) offers as evidence the lack of transparency in the structure and meaning of intonation, and Ladd (2008) notes that the paralinguistic and emotion-based characteristics often associated with intonation cause many to see it as a nonlinguistic feature in nontonal languages.
However, the autosegmental-metrical (AM) approach, developed by Pierrehumbert (1980), has paved the way for a number of studies that deal with Spanish intonation as a linguistic characteristic of the language (e.g. Face 2002, Hualde 2003, Nibert 2000, Prieto 2004, Sosa 1999). Although these studies deal with Spanish as a native language, the AM framework used in them can be extended to analyze nonnative Spanish as well.
This study examines the phonetic characteristics of certain intonational events in the speech of nonnative adult learners of Spanish who are involved in romantic relationships with native Spanish speakers. Spanish broad focus declaratives produced by these nonnative speakers were compared to the same sentences as produced by the participants' native speaking partners. Additionally, the nonnative Spanish sentences were compared to English broad focus declaratives as produced by the same normative participants. The intent of this study is to determine if acquisition of Spanish broad focus declarative intonation by native English speakers in a particular social context is possible, and, if it is possible, which factors contribute to it.
2. LITERATURE REVIEW.
2.1. ACQUISITION OF INTONATION. According to Nibert (2005), the acquisition of intonation by adult language learners is one of the least studied areas of second language acquisition research. This is unfortunate, as intonation is one of the most challenging linguistic features for adults to acquire and one of the features most essential to the production and comprehension of the appropriate message.
In terms of first language acquisition, intonation is one of the earliest features to be acquired by children. As a result, it quickly becomes automatic in speech, and native speakers find it nearly impossible to consciously describe its properties (Ramirez Verdugo 2006). If adults find it so challenging to define the characteristics of their own native intonation, it is to be expected that it would be that much more difficult for them to consciously learn the intonation of a nonnative language. Compounding the issue is that, at least in the United States, second language education is carried out principally by means of Communicative Language Teaching. While this method has much to recommend it, it does not overtly teach intonation; therefore, adults receive little to no explicit instruction in that area (Nibert 2005).
Despite the relative difficulty in learning second language intonation, its successful acquisition is essential to effective communication. Intonation is necessary in the organization, production and comprehension of messages. Nonnative speakers who use appropriate intonation in their second language are judged by native speakers to be better and more proficient than learners who do not use appropriate intonation (Ramirez Verdugo 2006, Tanner & Landon 2009). Failing to use appropriate intonation can lead to misunderstandings, communicative breakdowns and negative stereotypes of the second language speaker (Holden & Hogan 1993, Mennen 2007, Nibert 2005, Ramirez Verdugo 2006).
2.2. ACQUISITION OF SPANISH INTONATION. The relative paucity of studies that deal with adult acquisition of Spanish intonation is due in large part to the difficulties inherent in carrying out such work. The study of nonnative Spanish intonation is particularly challenging because native Spanish varies greatly across geographic regions (e.g. Prieto & Roseano 2010). It can therefore be difficult to successfully identity the intonation target of a particular learner, in particular if the learner lives in the United States and is exposed to a multitude of Spanish dialects. (1)
Several studies have resolved this difficulty by examining learners who lived and/or studied in a Spanish-speaking country. In such cases, the target may be understood to be that of the region in which the learners resided (e.g. Henriksen et al. 2010, Kelm 1987, Trimble 2013a, 2013b). While Kelm (1987) finds that transfer from English still affects his participants' Spanish intonation, the results from Henriksen et al. (2010) and Trimble (2013a, 2013b) show that their learners' intonation does indeed show movement towards the native target.
A possible reason for the difference in results may be that Kelm's (1987) participants consisted of missionaries, while the other studies involved students in study abroad programs. It can be assumed that both the missionaries and the students were motivated to use and understand the target language; such motivation may have positive effects on intonation acquisition, though its precise effects remain understudied (e.g. Ramirez Verdugo 2006, Tanner & Landon 2009). However, the ultimate goals of missionaries and study abroad students as residents in the country are not likely to be the same. The students are more likely to make learning the target language their specific goal and perhaps more likely to notice linguistic features of the language.
In any language learning environment, the learners' perception and production of the linguistic target play an important role. Major (1998) finds that the two processes are not mirror images of each other; however, there is a relation between them in the phonology of a learner's interlanguage. The interlanguage (Selinker 1972) refers to the linguistic system created by learners that contains elements taken from the native language and the target language, as well as elements common to neither language, such as linguistic universal (e.g. Major 1998, Tarone 1972). Major (1987) proposes that adult phonological acquisition proceeds via the ONTOGENY MODEL, in which both native language interference and developmental processes affect learners' speech as they increase in proficiency. Tarone (1972), Ramirez Verdugo (2006) and Tanner and Landon (2009) suggest that explicit instruction, learner awareness and conscious imitation of native speakers may aid in the acquisition of intonation. The learners' accurate perception of the target assists in the development of their interlanguage and the improvement of their linguistic production. This may have been the case with the learners in the study abroad programs.
Both Nibert (2005) and Trimble (2013a) find that learners with more experience improve in their perception of Spanish intonation. Trimble (2013b) conducted an additional production study that included learners from the same Venezuelan study abroad program that was used in Trimble (2013a). He found that both perception and production of dialect-specific intonation improve over time as the learners gain greater proficiency. In addition to these factors, the learners who produced the most consistent dialect-specific patterns by the end of the students' time abroad were those who had developed the strongest relationships with native speakers. Perception and production, on the basis of Trimble's (2013a, 2013b) work, do affect each other, particularly in contexts in which learners and native speakers frequently interact.
Such native-nonnative communication has been modeled in interactionist theories in second language acquisition research. These perspectives model language learning as a process that is influenced by learners' interactions and social closeness with native speakers (e.g. Larsen-Freeman 1991). Communication accommodation theory (Giles et al. 1977), the idea that speakers adapt features of their speech to more closely approximate those of their interlocutors, may also play a role in the results observed by Trimble (2013b).
The question that remains, however, is if learners make such gains when not in a study abroad context. Are learners in the United States who interact frequently with native speakers of Spanish capable of acquiring dialect-specific intonational targets? Do they accommodate their speech to the native speakers with whom they interact? To answer these questions, the learner population selected for this study consisted of normative adult learners of Spanish involved in romantic relationships with native Spanish speakers. Such a context would imply both frequency of communication and social closeness in the native-nonnative speaker interactions.
2.3. THE AUTOSEGMENTAL-METRICAL APPROACH. The AM approach (Pierrehumbert 1980) is a means of using discrete units to categorize a continuous intonation contour. Intonation contours are divided into tones that are associated with certain elements in the utterance. As this paper deals exclusively with pitch accents, the explanation of the AM approach will be limited to this feature.
Pitch accents are the international movements associated with stressed syllables in an utterance. They mark the point at which the intonation contour changes direction, thereby being categorized as high tones (peaks) and low tones (valleys). Pitch accents are also categorized in terms of the pitch accents' positions relative to the stressed syllable; the pitch accent can be tonic or posttonic, in that the direction change occurs, respectively, either within or after the stressed syllable. Changes based on the pitch accents' height and alignment can express different sentence types and pragmatic values, and particular height and alignment properties are also associated with different dialects (e.g. Face 2002, Prieto 2004, Sosa 1999, Willis 2003).
The AM model was originally applied to English intonation (Pierrehumbert 1980) and modified into the ToBI (Tones and Break Indices) transcription system (e.g. Silverman et al. 1992). The system has been modified for Spanish, resulting in Spanish ToBI, or Sp_ToBI (e.g. Beckman et al. 2002). The Sp_ToBI training website (Aguilar et al. 2009) provides explanatory materials and examples of particular pitch accents, as shown in figure 1. The shaded area represents the stressed syllable, while the line shows the direction of the intonation contour.
The work that has been done on the production of native Spanish intonation has tended to focus on several metrics. For instance, a number of studies have investigated peak and valley measurements, examining their relative height and their temporal alignment (e.g. Face 2008, Prieto 1998, 2004, Prieto et al. 1996, Prieto et al. 1995, Willis 2003). The height of peaks and valleys is examined to see if downstepping (lowered pitch) of successive pitch accents occurs, and alignment is measured to determine where the peaks and valleys occur with respect to stressed syllable boundaries. These investigations have also considered whether the stressed syllable contains a pitch accent or not. Deaccenting--the absence of a pitch accent--has been shown to correlate with certain types of speech, words, and sentence positions (e.g. Face 2003, Rao 2006, 2009). Other measurements that have been studied include stressed syllable duration, pitch intensity and the pitch accent rise (e.g. Face 2002, Willis 2002). These investigations have studied native Spanish intonation in a variety of contexts--declarative sentences, absolute interrogatives, pronominal interrogatives, broad focus, narrow focus, contrastive focus, read speech and spontaneous speech (e.g. Alvord 2006, Face 2002, 2003, 2008, Henriksen 2010, Rao 2006, 2009, Sosa 2003, Willis 2003).
In terms of broad focus declarative sentences (i.e. statements made in which all lexically stressed words in the sentence receive equal focus), there are a number of similarities between most dialects of native Spanish and native American English. For instance, declaratives are characterized by a final fall in intonation (e.g. Astruc et al. 2010, Face 2004, 2008, Liu & Xu 2007, Pierrehumbert & Hirschberg 1990, Willis 2005). The pitch accents may differ in terms of their alignment. In Spanish, the peak of the final pitch accent aligns within the tonic syllable (H* or L+H*, using the notations in figure 1), while earlier peaks align with posttonic syllables (L+>H*). This is fairly consistent across dialects (e.g. Astruc et al. 2010, Face 2008, Willis 2003, 2005). The generalizations that have been made regarding the intonation of American English tends to be based on speakers from the American Midwest (e.g. Liu 2009), and their declaratives tend to have peaks that are all aligned within the stressed syllable (H* or L+H*) (e.g. Liu & Xu 2007, Pierrehumbert & Hirschberg 1990).
Although the AM approach has been used principally in research on native languages, it has also been applied to second language research (Nibert 2005, Trimble 2013a, 2013b). The principal benefit is that AM research done across different languages and dialects, with both native and normative speakers, will yield results that are more standardized and can be more easily compared (Mermen 2007). Thus, the AM framework was selected for this study.
3.1. THE PARTICIPANTS. The focus of this study was to analyze the Spanish intonation contours produced by native speakers of American English (NSEs). Given the dialectal variation inherent in Spanish intonation and the difficulty of finding an adult learner in the United States who had only ever been exposed to one dialect, the participants chosen were NSEs involved in romantic relationships with native speakers of Spanish (NSSs). Assuming a high likelihood of intimacy and regular communication in Spanish within the context of the relationship, the specific dialects of the NSSs may be assumed to be the acquisition targets of the corresponding NSEs.
Five NSE/NSS relationships were selected for this study. All NSEs were raised in the American Midwest and lived in Minnesota at the time of the recording. Their romantic partners were NSSs who had been born outside of the United States or raised in a Spanish-speaking household by parents born outside of the United States. The NSEs completed a background questionnaire, based in part on Potowski (2004) and Potowski & Matts (2008), to provide information about their use of Spanish, their relationship, how Spanish is used in the context of the relationship and any observations and opinions of their partners' specific dialects. This was done to determine the amount of daily exposure that the NSEs have to their partners' dialects, as well as their level of commitment to that particular dialect, as Tanner and Landon (2009) observe that the effect of country-specific attitudes on intonation is an area that remains to be studied. The questionnaire from this study is included in Appendix A.
The NSEs' proficiency in Spanish was not formally assessed; however, at the time of the recording, they completed an additional, unanalyzed task in which they spoke spontaneously in Spanish with their partners. These recordings were assessed by the researcher, and the ACTFL Proficiency Guidelines (ACTFL 2012) were used to informally evaluate the participants' speech and gain a general idea of their proficiency.
The participants' information is given in table 1. (2)
Observing table 1, one may notice a difference between NSE1 and the other four participants. NSE1 was the only participant whose proficiency was assessed to be at the novice level, and she had studied Spanish for considerably less time than the other NSEs. She was also the only participant whose relationship had begun before her study of Spanish. It is therefore possible that her target was more likely to correspond with NSS1's particular dialect. Unlike the other participants, she had not been exposed to other potential targets before meeting her partner.
On the other hand, NSE1's commitment to Spanish in general and to NSS1's dialect in particular was weaker than that of the other participants. All participants expressed the belief that speaking Spanish was important to them. However, the other participants, in particular NSE4, far outpaced NSE1 in terms of daily Spanish use, which can be seen in Table 2.
NSE2, NSE3 and NSE5 all reported speaking equal amounts of English and Spanish with their partners. The other two participants showed a clear preference for one language; NSE4 used primarily Spanish, and NSE1 used primarily English. (5) There seems to have been a lack of consistency between NSE1's beliefs and her daily activities, perhaps due to her lower proficiency level.
The speakers all recounted Spanish-language interactions with members of their partners' families. NSE1 used Spanish, admittedly not often, with NSS1's grandmother, NSE4 spoke with NSS4's parents once a month and NSE2, NSE3 and NSE5 spoke several times a week with their partners' parents. NSE2 and NSS2, the only couple who were parents at the time of the recording, reported speaking in Spanish with their two children. Additionally, NSE3, NSE4 and NSE5 traveled somewhat regularly to Mexico to visit their partners' families.
The participants were asked various questions to probe their metalinguistic knowledge of Spanish dialects. NSE3, NSE4 and NSE5 reported being able to identify particular characteristics of their partners' dialects. NSE3 and NSE4 both mentioned phonological characteristics, and NSE3 specifically stated that he could recognize NSS3's intonation. When asked which dialect they believed to be the best variety of Spanish, NSE3 and NSE4 wrote that they believed that their partners' Mexican dialect was the best. NSE2 believed it to be Peninsular Spanish, since the language originated there, and NSE5 believed it to be Guatemalan Spanish. Having spent tune in Guatemala, she was particularly fond of that dialect and reported being able to identify its lexical characteristics, in addition to those of her husband's Mexican dialect.
The participants' professions must also be considered. NSE1 and NSE2 were, at the time of the recording, undergraduate students enrolled in Spanish classes at a large Midwestern university. NSE4 was a graduate student in Spanish at the same university; in addition to taking classes in Spanish, she also taught the language. Both NSE3 and NSE5 were high school teachers; NSE3 taught ESL and NSE5 taught Spanish. The participants' professions entailed regular exposure to Spanish and to potential targets aside from their partners' dialects.
3.2. THE TASK. The experiment consisted of tasks in which the speakers read English and Spanish broad focus declarative sentences. Broad focus declarative sentences were selected because they have been analyzed in the majority of AM intonation research, and they are considered to be the default sentence type (Face 2003).
Two sets of utterances were created for this task--Spanish declaratives and English declaratives. The Spanish declaratives used in this study were largely based on those used by Willis (2005) in his study on Mexican Spanish. All Spanish and English sentences contained three lexically stressed words, each of which was stressed on the penultimate syllable. The stress was placed on the penultimate syllable to allow enough wordspace for any delayed pitch accents to be realized. The utterances that were used in this study are included in Appendix B.
The utterances that were the focus of the study were the Spanish declaratives produced by the NSEs. The NSEs read the ten declarative utterances three times, for a total of 30 Spanish utterances (10 sentences x 3 readings). As in Kelm (1987), two control groups were used. The first group consisted of the English declaratives produced by the same NSEs, for the purpose of establishing the NSEs' native intonation patterns. These ten sentences were read three times, for a total of 30 English control utterances. The second control group consisted of the same Spanish declaratives produced by the NSS partners, for the purpose of establishing the intonation patterns specific to the dialect spoken in each relationship. These ten sentences were read three times, for a total of 30 Spanish control utterances. A Marantz PMD 660 recorder was used to record the utterances, and the recordings were conducted in a quiet room.
3.3. DATA MEASUREMENTS. The features selected to be measured in this study were the alignment and height of the peaks and valleys of each of the three pitch accents. As indicated in section 2.3, there are a number of other intonation characteristics that could have been analyzed. However, peak and valley position were selected to limit the scope of the project to the features that are principally examined in AM research and to provide a general idea of the intonation contour in an understudied type of Spanish (e.g. Alvord 2006, Elordieta 2003, Willis 2003).
Any disfluencies in the sentences read by the participants were discarded. For each of the remaining utterances, the presence and/or absence of the initial, medial and final pitch accent was observed. The absence of a pitch accent, or deaccenting, was defined using Willis' (2002) criterion of a movement of 7 Hertz (Hz) or less in the intonation of the stressed syllable. Praat (Boersma & Weenink 2011) was used to identify the syllable boundaries and the high and low tones in each utterance. Peak height and valley depth for the three pitch accents were measured, and the scaling change between each of the pitch accents was calculated to determine if downstepping occurred. Kelm (1995) proposes limiting participants to all male or all female speakers in order to avoid sex-based differences in pitch range. However, this was impossible, given that the participants in this study were all involved in heterosexual relationships. (6) Therefore, the scaling values from Praat were converted from Hz to the Equivalent Rectangular Bandwidth (ERB) scale. According to Colantoni & Gurlekian (2004), this scale standardizes pitch movements made in different registers and allows comparisons between male and female speakers. The alignment of the peaks and valleys of each pitch accent was also calculated, from the offset and the onset of the stressed syllable, respectively.
In order to determine the intonation patterns specific to the different speakers and languages, the average of the 30 measurements for each of the criteria was taken. This was assumed to represent the speakers' declarative intonation pattern, although it must be noted that there is a degree of variation in patterns across utterances and contexts, and that read speech is not necessarily representative of patterns found in spontaneous speech (e.g. Face 2003, Rao 2006, 2009).
Paired-samples t-tests were run to compare the NSEs' Spanish declaratives with both control sets. As the purpose of the paired t-test is to detect differences between two groups of matched pairs, the failure to reject the null hypothesis (achieving a p-value higher than alpha, which was assumed to be 0.10) would indicate that the two groups are similar. (7) The t-tests were run to determine if there were any significant differences in peak and valley scaling and alignment between the NSEs' Spanish declaratives and the Spanish (NSSs' Spanish declaratives) and the English (NSEs' English declaratives) control sets.
4. Results. Table 3 shows the number of declaratives produced by each of the ten participants, out of 30 total. In table 3, and in other tables in this paper, the set of sentences under study is in the white rows, while the two control groups are listed below it in the gray rows.
While most of the participants successfully produced most, if not all, of the 30 declaratives, there were two who struggled with the task. Both NSE1 and NSS5 produced a fairly low number of Spanish declaratives--22 and 23, respectively. NSE1's low number of fluent utterances is possibly due to her low proficiency in Spanish, but it is not surprising that any speaker should find the task challenging. Being recorded while reading sentences without any pragmatic context does not approximate normal communication, and speakers may react negatively to being asked to perform such a task (e.g. Barnes & Michnowicz 2013).
Of the sentences that were successfully produced, the percentage of deaccenting on the three stressed words, the alignment of the peaks and valleys and the change in peak and valley scaling were measured.
4.1. DEACCENTING. Table 4 shows the percentage of initial, medial and final stressed syllables that did not receive a pitch accent, or had any intonational movement of less than 7 Hz.
Although there are a number of different deaccenting patterns represented, several commonalities can be observed. First, for all speakers and both languages, the initial stressed syllable was almost never deaccented. In general, the final syllable was the one most likely to be deaccented. Although Face (2003) finds that deaccenting occurs more frequently in spontaneous speech than in read speech, only one speaker (NSE2 in her Spanish sentences) did not deaccent any stressed syllables. Deaccenting was observed in all other sets of read sentences.
Other than these similarities, the NSEs' Spanish sentences do not seem to strongly resemble the Spanish spoken by their partners. NSE3's Spanish sentences are similar to his English sentences, in that the initial stressed syllables are never deaccented, and the medial stressed syllable is more likely than the final stressed syllable to be deaccented, although this deaccenting is not very likely. NSE4's Spanish also resembles her English, in that the initial syllable is never deaccented, the final syllable is always deaccented, and the medial syllable is occasionally deaccented. It must be noted that her Spanish also shows a certain similarity to NSS4's Spanish; however, his medial syllable is more likely to be deaccented than hers. The remaining NSEs present Spanish patterns that are unlike either control group.
4.2. PEAK/VALLEY SCALING CHANGE. The scaling change refers to the change in frequency between adjacent peaks and valleys. If lower values are observed throughout the sentence, downstepping has occurred.
The only scaling change that was measured is the change that occurred between the initial and medial pitch accents. This was done because at least one participant in every relationship always or almost always deaccented the final stressed syllable; therefore, no final scaling change was produced that could be measured and compared to his/her partner's speech.
Table 5 shows the average scaling change from the initial stressed syllable to the medial stressed syllable for all speakers and both languages. The values are given in ERBs and are equal to the height of the initial peak/valley minus the height of the medial peak/valley. Positive numbers, therefore, indicate a down-stepped medial peak/valley. The results of the t-tests are also shown, assuming an alpha level of 0.10. The p-value shown to the right of the mean and standard deviation reflect that particular control group's comparison with the NSE Spanish, and any differences that were significant are marked with an asterisk.
In general, across speakers and languages, peaks tended to be downstepped more than valleys did. Downstepping occurred far more frequently than did upstepping (raised pitch on successive pitch accents), and any upstepping that did occur was very minimal.
In terms of peak scaling, NSE1, NSE2 and NSE5 produced Spanish declaratives that were not significantly different from either control group. However, the peak downstepping observed in the declaratives produced by NSE1 and NSE2 is similar to that of their partners. Although firm conclusions cannot be drawn, this indicates a possible movement towards their partners' targets. The numbers observed for NSE5 show that her peak scaling was more similar to her own English than to NSS5's Spanish. Both NSE3 and NSE4 produced peaks whose scaling was significantly different from that of both control groups. In fact, the raw data shows that they seemed to have moved away from their partners' Spanish targets.
With respect to valley scaling, the participants show fewer significant differences from either control group. The one participant whose Spanish valleys were significantly different from her English ones is NSE2, who seemed to have approximated NSS2's valley scaling pattern. The other participants all showed movement towards their partners' targets, but there was no statistical significance, and definite conclusions cannot be drawn.
4.3. PEAK/VALLEY ALIGNMENT. Peak/valley alignment refers to the position of the peaks and valleys with respect to the stressed syllable offset and onset, respectively. Again, given the high percentage of deaccenting on the final stressed syllable, the only alignment values that were measured were those of the initial and medial stressed syllables.
Table 6 shows the average peak and valley alignment for the initial pitch accent. The values are given in milliseconds; a positive value indicates alignment after the offset (peaks) and onset (valleys), while a negative value indicates alignment before the offset/onset. The results of the t-tests are also shown, assuming an alpha level of 0.10. The p-value shown to the right of the mean and standard deviation reflect that particular control group's comparison with the NSE Spanish. An asterisk indicates a significant difference between the NSE Spanish utterances and the control group.
In terms of the initial pitch accent, almost all NSEs aligned the peak well into the post-tonic syllable, producing a pitch accent that can be classified as L+>[H.sup.*]. The exception is NSE3's initial peak, which, while it is still L+>[H.sup.*], occurs very soon after the offset of the stressed syllable. The NSSs all align the initial peak posttonically, some more so than others, likewise producing L+>[H.sup.*]. The same pitch accent is observed in the NSEs' English declaratives.
The NSE Spanish valleys are all aligned prior to the onset of the stressed syllable, although NSE1's valley occurs only just before the onset. The NSE Spanish valleys resemble the NSS Spanish valleys in this respect, with the exception of NSS2, whose valley occurs just after the onset. A striking difference is observed between the NSE Spanish valleys and the NSE English valleys. The latter sets of utterances are all aligned after the onset of the stressed syllable. While the Sp_ToBI training website (Aguilar et al. 2009) does not present pitch accent labels that differ in terms of valley alignment, Face (2011) proposes that the star notation be used to differentiate early and late valleys. The NSE and NSS Spanish declaratives with the early valley can be labeled with L, while the late valley of the English utterances can be labeled as [L.sup.*].
The statistical results reveal that there are significant differences between the NSE Spanish sentences and the control groups, despite the similarities observed with respect to pitch accent type. NSE3 and NSE5 produced pitch accents that were significantly different from either control group, while NSE4's pitch accent was significantly different only from her own English utterances. NSE1's Spanish resembled both control groups, diverging from her English pattern only with respect to valley alignment, while NSE2's production was nearly the opposite. Her Spanish resembled neither her English nor NSS2's Spanish, with the exception of her English-like peak alignment.
Table 7 shows the average peak and valley alignment for the medial pitch accent. Much like the initial peak, the medial peak for nearly all speakers was aligned posttonically. The only exception was NSS5, whose peak was aligned within the stressed syllable, just prior to the offset. Additionally, NSS2's peak was aligned posttonically, but it occurred very shortly after the offset.
The medial valleys were much more varied in their alignment. NSE1, NSE2 and NSE3 placed their Spanish valleys very close to the onset, while NSE4 and NSE5 had much later valleys. In their English declaratives, all valleys were late. NSS1, NSS2 and NSS3 had late valleys, NSS4's valley was aligned just prior to the onset, and NSS5 had the earliest medial valley of all participants.
The statistical results show that, again, NSE5 produced a medial accent that resembled neither control group. NSE1 and NSE2 nearly did the same; however, NSE1's peaks resembled her English pattern, while NSE2's valleys were not significantly different from NSS2's utterances. NSE3 and NSE4 produced similar declaratives that resembled the Spanish control group in peak alignment and the English control group in valley alignment.
5.1. RESULTS SUMMARY. The five NSE participants' Spanish intonation for broad focus declaratives was compared with 1) their own native English intonation and 2) their NSS partners' Spanish intonation to see if their second language intonation approximated the target or if it was affected by transfer from English. The intonation contours were analyzed according to the following criteria--the percentage of deaccenting, the intonation change between peaks, the intonation change between valleys, the alignment of the initial and medial peaks from the stressed offset and the alignment of the initial and the medial valleys from the stressed onset. The results were varied. Some NSEs seemed to approximate certain criteria of the target intonation, others showed transfer of their native patterns, and some used patterns that could be attributed to both control groups and some presented patterns unobserved in either control group.
Table 8 labels the initial, medial and final pitch accents produced by all of the participants. Although the final pitch accents were not analyzed in sections 4.2 and 4.3, due to the high percentage of deaccenting by some of the participants, measurements were taken in order to determine the characteristics of those that were produced. Combining the labeling conventions provided by Aguilar et al. (2004) and Face (2011), the labels used in this project are defined in (1). Down-stepping is indicated with the "!" notation, as mentioned in Beckman et al. (2002).
(1) a. L+[H.sup.*]--valley before onset, peak before offset
b. L+>[H.sup.*]--valley before onset, peak after offset
c. [L.sup.*]+H--valley after onset, peak before offset
d. [L.sup.*]+>H--valley after onset, peak after offset (8)
Nearly all speakers in both languages observed a distinction between the initial/medial pitch accents and the final pitch accent. The initial and medial pitch accents were much more likely to have posttonic peaks, while the final pitch accent, in the instances in which it was produced, contained peaks within the tonic syllable. This has been observed in native Spanish (e.g. Astruc et al. 2010, Face 2008, Willis 2003, 2005), but not necessarily in native English (e.g. Liu & Xu 2007, Pierrehumbert & Hirschberg 1990). The NSEs' nonnative Spanish likewise followed the same pattern, though they, with the exception of NSE4, were less likely to deaccent the final stressed word.
In terms of individual speakers and their partners' targets, the speakers' results are also varied. NSE1 matched NSS1 on the initial and medial pitch accents, but her final accent was different from either the target or her own English. NSE2 produced three pitch accent types that differed from either control group. Both NSE3 and NSE5 matched their partners on the initial accent, maintained their English patterns on the medial accent and produced a different pattern on the final accent. NSE4's results were similar; she matched NSS4 on the initial accent and maintained her English pattern on the medial accent. Her final stressed word received no pitch accent, which matched both control groups.
The statistical results present a slightly different picture, as they examine more closely the exactness of the particular peak/valley height and alignment patterns. Table 9 summarizes these results for each intonational characteristic, based on the two control groups. An X indicates that what was observed for that particular speaker's characteristic was not similar to the control group. A V indicates that the characteristic was similar to the control group. The p-values were used to indicate which of the two symbols was placed in the appropriate cell.
Some similarities can be noted in the speakers' results. All five participants produced outputs that did not resemble either control group, and all five also produced patterns that were not significantly different from the two control groups. This seems to be indicative of the effect of interlanguage, in that elements both common and foreign to the native and normative languages were found in their speech. NSE1, NSE2, NSE3 and NSE4 produced certain features that were similar to their partners' Spanish, indicating possible acquisition, as well as other features similar to their own English, indicating possible transfer.
5.2. Acquisition of broad focus declaratives. This study sought to determine if nonnative speakers of Spanish could acquire dialect-specific broad focus declarative intonation in a non-study abroad setting and, if so, which factors had an effect on their acquisition. The participants selected differed in terms of their proficiency levels, their exposure to their partners' dialects of Spanish, their exposure to other varieties of Spanish, their daily use of Spanish, their travels, their professions, and the duration of their relationships.
A particularly notable divide, as has been mentioned, was that between NSE1 and the other participants. NSE1 had had little exposure to Spanish, her proficiency was low, and she did not use the language with any serious frequency. Given these facts, it would be expected that her Spanish intonation of broad focus declaratives would more strongly resemble her English patterns. However, the results show that she produced declaratives that did, on certain criteria, resemble NSS1's utterances. In this way, she did not differ greatly from the other NSE participants.
However, another difference between NSE1 and the other participants has to do with when they started learning Spanish. NSE1 was the only participant who had met her partner and begun her relationship before beginning her formal study of Spanish. It is possible that NSE1, despite not using much Spanish with NSS1, had been strongly influenced by her partner's dialect, the first with which she had felt a significant connection. The other participants, on the other hand, had studied Spanish for years before meeting their partners, and it is possible that their intonation patterns had already been set. There may have been little to no correlation between the time spent using Spanish in the relationship and the acquisition of broad focus declarative intonation. Perhaps second language intonation, like native intonation, is acquired early and is thus based on targets with which the normative speakers had first significant contact.
It must be noted that early exposure to the dialect used in the relationship might not have been the only influence on the acquisition of broad focus declaratives. The other speakers seemed to have acquired certain characteristics of their partners' dialects. These speakers were all highly proficient in Spanish, and they used it regularly with their partners and their partners' families. NSE4, however, stands out from the other participants. It might be expected that her results would be similar to NSE3 and NSE5, based on their proficiency levels, their travels and their metalinguistic knowledge of their partners' dialects, but NSE4 shows a greater acquisition of her partner's intonation. According to the information that she provided, she and NSS4 were the only couple in the study who use Spanish as their primary household language. Perhaps her favoring Spanish over English in daily interactions with her fiance allowed NSE4 to acquire his declarative pattern six years after having begun her study of Spanish and despite working in an environment with Spanish speakers of multiple dialects.
Finally, mention should be made of NSE5's pattern. Her pitch accent alignment and deaccenting pattern resembled neither her own English nor NSS5's Spanish. According to the information that she provided, it was expected that she would behave like NSE3. Both participants used Spanish and English equally within their relationship, used Spanish to communicate with their partners' families and had traveled to their partners' countries. However, NSE3 showed greater acquisition of his partner's intonation than did NSE5.
NSE5, therefore, did not appear to have accommodated to her partner's dialect. Research on communication accommodation has found that accommodation is less likely to occur when speakers wish to establish boundaries between members of different groups (e.g. Callahan 2006, Heller 1982, Lo 1999, Weyers 1999). Might NSE5 have wished to separate herself from NSS5's Mexican dialect? In her answers on the questionnaire, she placed a strong emphasis on Guatemala. In her response to how frequently she uses Spanish media, she stated, 'Not very often--when I'm curious about what's going on in Guatemala or any other Spanish-speaking country.' It is worth noting the distinction that she makes between Guatemala--important enough to be specified--and other Spanish-speaking countries--less important and lumped together in a group. In her response to whether or not she could identify the characteristics of her partner's dialect, she wrote, 'We were talking one day on the phone, and he didn't understand [the word I used] ... I was using my Guatemalan dialect. Since then, my dialect has become more similar to his, although I make sure to remember the Guatemalan way as well.' In this, she noted her active and ongoing commitment to Guatemalan Spanish. Finally, she has spent a total of three months in Guatemala, versus one and a half months in Mexico. Perhaps her lack of accommodation to her partner's Mexican dialect was a means of maintaining a previously-established identity with Guatemalan Spanish.
This indicates the importance of clearly defining the particular target of nonnative speakers. It is quite possible that NSE5's target was Guatemalan Spanish, rather than her husband's own Mexican variety. Unfortunately, Guatemalan Spanish intonation is not a popular topic of investigation and does not appear in the literature. Given this lack of information, there was no way to compare NSE5's production with any Guatemalan broad focus declaratives. Additionally, the intonation contour that she produced (posttonic initial and medial peaks, followed by a tonic final peak) is a contour common to many dialects. In Prieto and Roseano's (2010) edited collection of Spanish intonation, this particular contour is noted in Castilian, Venezuelan Andean and Chilean Spanish. Without a clearer understanding of NSE5's target intonation (Guatemalan Spanish or any other variety to which she had been exposed during 17 years speaking Spanish), firm conclusions cannot be made. Rather, the only real conclusion is that, despite her fluency in Spanish and her regular use of it within the context of her relationship, she did not show accommodation to her partner's intonation.
The results show that nonnative speakers' acquisition of certain features of their partners' intonation seems to be dependent on several factors. The partners' dialect must be the speakers' target, which can be challenging for speakers who had spoken Spanish for years before meeting their partners. They may have already acquired the intonation of another dialect, whether unconsciously through early exposure or consciously as in NSE5's indicated preference for Guatemalan Spanish over Mexican Spanish.
Regular and frequent use of the language within the context of the relationship seems to also be a necessary aspect of acquiring a particular intonational pattern. Of the four participants who had been exposed to Spanish for years before their relationship began, only NSE4 used Spanish as her primary language of communication with her fiance. Her declaratives likewise showed a greater similarity to her partner's target than did those of the other nonnative participants. Therefore, it seems likely to conclude that the acquisition of dialect-specific intonation patterns does occur, to a certain degree and under certain circumstances, in the context of a romantic relationship with a native Spanish speaker.
5.3. APPERCEIVED INPUT. In this study, the participants' proficiency level of Spanish was shown to matter less in the acquisition of dialect-specific intonation than their identification of the target, their motivation to use that target and the frequency with which they spoke Spanish. Gass' (1988) integrated framework of second language acquisition provides an interpretation of these findings.
Gass' (1988) model of acquisition consists of a process with 5 stages: 1. apperceived input, 2. comprehended input, 3. intake, 4. integration and 5. output. The first stage, apperceived input, deals with the language to which nonnative speakers are exposed and the input that they receive. Gass and Selinker (2008) present several factors related to apperceived input that may affect what normative speakers are able to notice, some of which are relevant to this study.
The first relevant factor is frequency. A feature that is more frequent in the input is more likely to be noticed by learners. This particular factor helps to explain NSE4's intonation. As the only participant who reported using Spanish as her primary language of communication in the relationship, she would have received greater frequency of input of her partner's specific intonation pattern than the other NSEs.
The second relevant factor is affect, which includes social distance, status, motivation, and attitude. If learners feel distant from the target language community, there will be little to no input available to them. While this is obviously not the case in this study, considering that the participants were in relationships with speakers from the target language community, it may relate to NSE5's nonaccommodation of her partner's dialect. She may not have felt socially distant from it, but she may have felt a greater connection to a different dialect. She may have had stronger motivation to continue to use what may have been Guatemalan Spanish than to switch to the Mexican Spanish of her partner.
The final relevant factors are attention and prior knowledge. The learners must attend to the input in order to notice a mismatch between their prior knowledge of the second language and the language as it is actually produced by native speakers. This may help to explain why NSE1 showed acquisition similar to that of NSE2 and NSE3, despite the great disparity in their proficiency levels and time spent studying Spanish. Intonation is not generally transparent (e.g. Nibert 2005), and NSE2 and NSE3 might be less likely to notice any mismatches between their own intonation and that of their native partners. However, NSE1 had had no experience with Spanish before beginning her relationship with NSS1. This lack of prior knowledge meant that she had to pay attention to NSSl's Spanish; it was the only input that she had before beginning Spanish classes six months later.
Although the factors listed by Gass and Selinker (2008) refer to the acquisition of nonnative grammar, the idea of apperceived input can also be applied to the acquisition of nonnative intonation. Learners must notice the relevant and frequently-occurring features in the input, they must be able to comprehend them and fit them into their already-established system, and they must feel a connection to the target language community. The learners in this study who met these characteristics were more likely to more closely approximate the dialect-specific intonation of their romantic partners.
5.4. STUDY LIMITATIONS. The results of this investigation must be viewed as a starting point in the study of normative Spanish intonation. Several features of one type of read sentence were analyzed, and the study consisted of a very small number of participants. More studies need to be done in the field of nonnative acquisition of Spanish intonation. These studies should continue to examine the characteristics of normative broad focus declaratives, to determine if the same factors affect intonation acquisition, as well as extend the investigation to other types of read sentences. For instance, there is a great deal of dialectal variation in the intonation contours of absolute interrogatives (e.g. Alvord 2010, Face 2004,2008, Willis 2004, 2005), and a study of this type of utterance might shed more light on NSEs' accommodation of their partners' dialects. Further research should also consider spontaneous speech, to see if any intonation characteristics discovered in read speech transfer to interactions in more natural environments. The speech of other NSE/NSS couples should also be examined. Future studies should include a greater number of participants to determine if the results observed in this study are supported in the contexts of other relationships with possibly different characteristics. Such future participants might include NSEs and NSSs of the same sex, in order to avoid any issues relating to gendered intonation differences.
With regard to the specifics of this study, the original intent was to analyze boundary tones (the intonation contours associated with the end of the utterance) as well as pitch accents. However, the only speakers who produced usable boundary tones were NSE4 and NSS4. NSE1, NSS1, NSE2 and NSE5 produced a final rise, rather than a final fall, in the majority of their utterances. It is possible that this contour is not due as much to the speakers' declarative speech patterns as to a list effect created by their reading a series of utterances. Nibert (2000) observed similar results from her own participants in a comparable task. Given these results, the final contours were not analyzed. Future studies would do well to devise a task that negates the list effect. Additionally, the deaccenting observed in the speech of the participants should be more closely examined, to determine if it actually is deaccenting, which is uncommon in read speech, or if it involves final lowering. In the case of final lowering, a pitch accent would still be perceived by listeners (e.g. Face 2003, Prieto et al. 1996).
Finally, the ideal study of intonation acquisition in the context of a romantic relationship would record the initial state prior to any sort of language contact between the two participants. In the first place, English utterances from the NSE participants before their having begun the relationship should be recorded. Only in this way would it be possible to determine that the L2 has had no effect on native L1 patterns. If the NSSs' Spanish has affected the intonation of the NSEs' English, the entire English control group would be called into question. In the second place, Spanish utterances from the NSS participants before their having been exposed to the primarily English-speaking environment of the United States should also be recorded. If the NSSs' Spanish had been affected by their daily exposure to English, the Spanish control group would also be called into question. Obviously, the logistical details of such a study make it challenging to conduct, but the possibility of backwards interference make it a worthwhile proposition to consider.
6. CONCLUSIONS. There has been little research done up to this point on the acquisition of Spanish intonation by adult second language learners. This study attempts to add to the field by examining the Spanish of five native speakers of English who are involved in romantic relationships with native speakers of Spanish. The participants in the study produced Spanish broad focus declaratives, which were then compared to their own English declaratives as well as to Spanish declaratives produced by their partners. The speakers chosen for the study differed in terms of their proficiency and use of Spanish within the context of the relationship.
The results show that lower levels of Spanish proficiency do not seem to be an impediment to acquiring the dialect-specific intonation of broad focus declaratives. Rather, successful acquisition seems dependent on the identification of the appropriate target. If the speaker has extensive prior experience with Spanish, particularly prior experience to which he/she feels a strong emotional connection, the intonational target may not be the same as the dialect spoken within the relationship. On the other hand, if the relationship is the speaker's initial introduction to Spanish, it seems more likely that that dialect will be the speaker's target. In addition to the appropriate identification of the target, further acquisition of intonation seems dependent upon the speaker's commitment to speaking primarily in Spanish. If the speaker uses more English, or an equivalent amount of English and Spanish, he/she may not acquire as many target-like intonational features as a speaker who speaks mostly Spanish within the context of the romantic relationship. These results support the importance of apperceived input in the acquisition of intonation phonology.
This study, as previously stated, is an initial attempt to understand and analyze normative Spanish intonation. While efforts have been made to classify the intonation of students in a study abroad context, there has been little work done on the acquisition of Spanish intonation in a foreign language environment. The results provide a point of departure for future studies, which would be well served by expanding this work to analyze other features of declarative sentences, other types of read sentences and spontaneous speech. A larger number of participants should be included in future studies, to attempt to determine if the factors uncovered in this small-scale study do, in fact, affect the acquisition of dialect-specific intonation. It remains to be seen whether or not the characteristics of nonnative acquisition of Spanish intonation are consistent across speech types, dialects and social contexts.
Appendix A: Questionnaire
Please answer all that apply and clarify when necessary.
3. Where were you born?
4. How long have you lived in the Twin Cities area? II.
II. Language background
1. How long have you studied Spanish?
2. Have you traveled to any Spanish-speaking countries?
a. If so, where did you go and for how long?
3. Do you listen to/watch Spanish media?
a. If so, what and how often?
4. Do you read Spanish media?
a. If so, what and how often?
5. Why did you decide to learn Spanish?
6. How would you describe the Spanish language?
7. Have you found Spanish to be useful in your life?
a. Why/why not?
8. Where do you think the best variety of Spanish is spoken?
9. Do you speak any languages other than Spanish?
III. Relationship background
1. What is the status of your relationship (dating, married, etc.)?
2. How long have you been involved with your partner?
3a. If your partner is from another country:
a. What city/country is (s)he from?
3b. If your partner is from the United States:
a. Where are his/her parents from?
4. Where did you meet?
IV. Language and the relationship
1. What language do you speak with your partner? (Please select an answer)
a. All Spanish
b. Mostly Spanish
c. Spanish and English equally
d. Mostly English
e. All English
2. Please provide a numeric value for the following sentences.
a. On an average day, I speak in Spanish___% of the time with my partner.
b. On an average day, I speak in Spanish___hours and___minutes with my partner.
3. Are you in contact with Spanish-speaking members of your partner's family?
a. If so, with whom do you speak in Spanish?
b. How often do you speak with them?
c. How do you talk (in person, phone, email, Facebook etc.)?
d. Have they come to the Twin Cities?
i. If so, when and for how long?
4. Have you traveled to your partner's city and/or country?
a. If so, where and for how long?
5a. If you do not have children:
a. If you have children, will they know Spanish as well as you do?
i. Why/why not?
5b. If you have children:
a. What language(s) do y ou speak to them?
b. What language(s) does your partner speak to them?
c. What language(s) do they speak to you?
d. What language(s) do they speak to your partner?
e. When they are in their 20s, will they know Spanish as well as you do?
i. Why/why not?
f. Have they ever visited your partner's country?
i. If so, when and for how long?
6. Can you identify characteristics of your partner's dialect?
a. If so, what are they?
Appendix B: Utterances
Spanish broad focus declarative sentences for NSE and NSS:
1. Marina mira la luna 'Marina looks at the moon'
2. Mariana la vaba la lana 'Mariana washed the wool'
3. Elena amaba la mula 'Elena loved the mule'
4. La nina adora al minero 'The girl adores the miner'
5. Lenini me mima manana 'Lenini will spoil me tomorrow'
6. Maria lavaba la rana 'Maria washed the frog'
7. Lorena mimaba a la nina 'Lorena spoiled the girl'
8. La rana amaba a la nina 'The frog loved the girl'
9. Elena emula al nino 'Elena emulates the boy'
10. El nino miraba la luna 'The boy looked at the moon'
English broad focus declarative sentences for NSE:
1. You're selling some lilies tomorrow
2. Maria remembers the money
3. Your family envisioned me smaller
4. The lawyers are living in Miami
5. Warner is saving the money
6. Allie will marry the miner
7. Warren remembers Maria
8. Aaron has money for the families
9. Allie remembers the flowers
10. Mary has money for a lawyer
Meghann M. Peace
St. Mary's University
ACTFL. 2012. Proficiency Guidelines. Online: http://www.actfl.org/sites/default/ files/pdfs/ public/ACTFLProficiencyGuidelines2012_FINAL.pdf.
Aguilar, Lourdes; Carme De-la-mota; and Pilar Prieto. 2009. Sp_ToBI Training Materials. Online: http://prosodia.upf.edu/sp_tobi/.
Alvord, Scott M. 2006. Spanish intonation in contact: The case of Miami Cuban bilinguals. Minneapolis, MN: University of Minnesota dissertation.
Alvord, Scott M. 2010. Variation in Miami Cuban Spanish interrogative intonation. Hispania 93.2.235-55.
Astruc, Lluisa; Elsa Mora; and Simon Rew. 2010. Venezuelan Andean Spanish intonation. Transcription of the intonation of the Spanish language, ed. by Pilar Prieto and Paolo Roseano, 191-226. Munich: Lincom Europa.
Barnes, Hilary, and Jim Michnowicz. 2013. Peak alignment in Veneto-Spanish bilinguals. Paper presented at the annual meeting of the Hispanic Linguistics Symposium, Ottawa.
Beckman, Mary E.; Manuel DIaz-Campos; Julia Tevis McGory; and Terrell A. Morgan. 2002. Intonation across Spanish, in the Tones and Break Indices framework. Probus 14.9-36.
Boersma, Paul, and David Weenink. 2011. Praat: Doing phonetics by computer [Computer program]. Version 5.2.23. Online: http://www.praat.org/.
Callahan, Laura. 2006. English or Spanish?! Language accommodation in New York City service encounters. Intercultural Pragmatics 3.1.29-53.
Colantoni, Laura, and Jorge Gurlekian. 2004. Convergence and intonation: Historical evidence from Buenos Aires Spanish. Bilingualism 7.2.107-19.
Elordieta, Gorka. 2003. The Spanish intonation of speakers of a Basque pitch-accent dialect. Catalan Journal of Linguistics 2.67-95.
Face, Timothy L. 2002. Local intonational marking of Spanish contrastive focus. Probus 14.71-92.
Face, Timothy L. 2003. Intonation in Spanish declaratives: Differences between lab speech and spontaneous speech. Catalan Journal of Linguistics 2.115-31.
Face, Timothy L. 2004. The intonation of absolute interrogatives in Castilian Spanish. Southwest Journal of Linguistics 23.2.65-80.
Face, Timothy L. 2008. The intonation of Castilian Spanish declaratives and absolute interrogatives. Munich: Lincom Europa.
Face, Timothy L. 2011. Perception of Castilian Spanish intonation: Implications for intonational phonology'. Munich: Lincom Europa.
Gass, Susan. 1988. Integrating research areas: a framework for second language studies. Applied Linguistics 9.198-217.
Gass, Susan M., and Larry Selinker. 2008. Second language acquisition: An introductory course. New York: Routledge.
Giles, Howard; Richard Y. Bourhis; and Donald M. Taylor. 1977. Towards a theory of language in ethnic group relations. Language, ethnicity, and intergroup relations, ed. by Howard Giles, 307-48. London: Academic.
Heller, Monica S. 1982. Negotiations of language choice in Montreal. Language and social identity, ed. by John J. Gumperz, 108-18. Cambridge: Cambridge University Press.
Henriksen, Nicholas C. 2010. Question intonation in Manchego Peninsular Spanish. Bloomington, IN: Indiana University dissertation. Henriksen, Nicholas C.; Kimberly L. Geeslin; and Erik W. Willis. 2010. The development of L2 Spanish intonation during a study abroad immersion program in Leon, Spain: Global contours and final boundary movements. Studies in Hispanic and Lusophone Linguistics 3.1.113-162.
Holden, Kyril T., and John T. Hogan. 1993. The emotive impact of foreign intonation: An experiment in switching English and Russian intonation. Language and Speech 36.1.67-88.
Hualde, Jose I. 2003. El modelo metrico y autosegmental. Teorias de la entonacion, ed. by Pilar Prieto, 155-84. Barcelona: Ariel.
Kelm, Orlando R. 1987. An acoustic study on the differences of contrastive emphasis between native and non-native Spanish speakers. Hispania 70.3.627-633.
Kelm, Orlando R. 1995. Acoustic measurement of Spanish and English pitch contours: Native and non-native speakers. Hispanic Linguistics 6-7.435-448.
Ladd, D. Robert. 2008. Intonationalphonology. Cambridge: Cambridge University Press.
Larsen-Freeman, Diane. 1991. Second language acquisition research: Staking out the territory. TESOL Quarterly 25.2.315-50.
Lipski, John M. 2008. Varieties of Spanish in the United States. Washington, D.C.: Georgetown University Press.
Liu, Fang. 2009. Intonation systems of Mandarin and English: A functional approach. Chicago, IL: University of Chicago dissertation.
Liu, Fang, and Yi Xu. 2007. Question intonation as affected by word stress and focus in English. Proceedings of the 16th International Congress of Phonetic Sciences, ed. by Jurgen Trouvain and William J. Barry, 1189-92. Saarbrucken, Germany: Saarland University.
Lo, Adrienne. 1999. Codeswitching, speech community membership, and the construction of ethnic identity. Journal of Sociolinguistics 4.461-79.
Major, Roy C. 1987. A model for interlanguage phonology. Interlanguage phonology: The acquisition of a second language sound system, ed. by Georgette Ioup and Steven H. Weinberger, 101-25. Cambridge, MA: Newbury House.
Major, Roy C. 1998. Interlanguage phonetics and phonology. Studies in Second Language Acquisition 20.2.131-37.
Mennen, Ineke. 2007. Phonological and phonetic influences in non-native intonation. Non-native prosody: Phonetic description and teaching practice, ed. by iiirgen Trouvain and Ulrike Gut, 53-76. Berlin: Mouton de Gruyter.
Nibert, Holly. 2000. Phonetic and phonological evidence for intermediate phrasing in Spanish intonation. Champaign, IL: University of Illinois dissertation.
Nibert, Holly. 2005. The acquisition of the phrase accent by intermediate and advanced adult learners of Spanish as a second language. Selected proceedings of the 6th conference on the acquisition of Spanish and Portuguese as first and second languages, ed. by David Eddington, 108-22. Somerville, MA: Cascadilla Proceedings Project.
Pierrehumbert, Janet B. 1980. The phonetics and phonology of English intonation. Cambridge, MA: MIT dissertation.
Pierrehumbert, Janet, and Julia Hirschberg. 1990. The meaning of intonational contours in the interpretation of discourse. Intentions in communication, ed. by Philip R. Cohen, Jerry L. Morgan and Martha E. Pollack, 271-311. Cambridge, MA: MIT Press.
Potowski, Kim. 2004. Spanish language shift in Chicago. Southwest Journal of Linguistics 23.87-116.
Potowski, Kim, and Janine Matts. 2008. MexiRicans: Interethnic language and identity. Journal of Language, Identity and Education 7.137-160.
Prieto, Pilar. 1998. The scaling of the L tone line in Spanish downstepping contours. Journal of Phonetics 26.261-82.
Prieto, Pilar. 2004. The search for phonological targets in the tonal space: HI scaling and alignment in five sentence-types in Peninsular Spanish. Laboratory approaches to Spanish phonology, ed. by Timothy L. Face, 29-59. Berlin: Mouton de Gruyter.
Prieto, Pilar, and Paolo Roseano. 2010. Transcription of intonation of the Spanish language. Munich: Lincom Europa.
Prieto, Pilar; Chilin Shih; and Holly Nibert. 1996. Pitch downtrend in Spanish. Journal of Phonetics 24.4.445-73.
Prieto, Pilar; Jan van Santen; and Julia Hirschberg. 1995. Tonal alignment patterns in Spanish. Journal of Phonetics 23.429-51.
Ramirez Verdugo, M. Dolores. 2006. A study of intonation awareness and learning in non-native speakers of English. Language Awareness 15.3.141-57.
Rao, Rajiv. 2006. On intonation's relationship with pragmatic meaning in Spanish. Selected proceedings of the 5* Hispanic Linguistics Symposium, ed. by Timothy L. Face and Carol A. Klee, 103-15. Somerville, MA: Cascadilla Proceedings Project.
Rao, Rajiv. 2009. Deaccenting in spontaneous speech in Barcelona Spanish. Studies in Hispanic andLusophone Linguistics 2.1.31-75.
Selinker, Larry. 1972. Interlanguage. International Review of Applied Linguistics 10.209-31.
Silverman, Kim E. A.; Mary E. Beckman; John F. Pitrelli; Mari Ostendorf; Colin W. Wightman; Patti Price; Janet B. Pierrehumbert; and Julia Hirschberg. 1992. ToBI: A standard for labeling English prosody. ICSLP 92 Proceedings: International Conference on Spoken Language Processing, volume 2, 867-70.
Sosa, Juan M. 1999. Foneticay fonologia del espanol hispanoamericano. Amherst, MA: University of Massachusetts dissertation.
Sosa, Juan M. 2003. Wh-questions in Spanish: Meanings and configuration variability. Catalan Journal of Linguistics 2.229-47.
Tanner, Mark W., and Melissa M. Landon. 2009. The effects of computer-assisted pronunciation readings on ESL learners' use of pausing, stress, intonation, and overall comprehensibility. Language learning and technology 13.3.51-65.
Tarone, Elaine. 1972. Interlingual identification in pronunciation. TESOL Quarterly 6.4.325-31.
Trimble, John C. 2013a. Perceiving intonational cues in a foreign language: Perception of sentence type in two dialects of Spanish. Selected proceedings of the 15th Hispanic Linguistics Symposium, ed. by Chad Howe, Sarah E. Blackwell, and Margaret Lubbers Quesada, 78-92. Somerville, MA: Cascadilla Proceedings Project.
Trimble, John C. 2013b. Acquiring variable L2 Spanish intonation in a study abroad context. Minneapolis: University of Minnesota dissertation.
Utts, Jessica M., and Robert F. Heckard. 2006. Statistical ideas and methods. Belmont, CA: Thomson Brooks/Cole.
Weyers, Joseph R. 1999. Spanish as an ethnic marker in El Paso, Texas. Southwest Journal of Linguistics 18.1.103-16.
Willis, Erik W. 2002. Is there a Spanish imperative intonation revisited: Local considerations. Linguistics 40.2.347-74.
Willis, Erik W. 2003. The intonational system of Dominican Spanish: Findings and analysis. Champaign, IL: University of Illinois dissertation.
Willis, Erik W. 2004. Dominican Spanish absolute interrogatives in broad focus. Laboratory approaches to Spanish phonology, ed. by Timothy L. Face, 61-91. Berlin: Mouton de Gruyter.
Willis, Erik W. 2005. Tonal levels in Puebla Mexico Spanish declaratives and absolute interrogatives. Theoretical and experimental approaches to Romance linguistics, ed. by Randall Gess and Edward J. Rubin, 351-63. Amsterdam: Benjamins.
Department of Languages
St. Mary's University
San Antonio, TX 78228
(1) See Lipski (2008) for an overview of the different types of Spanish spoken within the United States.
(2) Understandable concerns might be raised regarding the couples' commitment to each other and, thus, the NSEs' commitment to learning, speaking and understanding their partners' dialects. The couples were selected based on the seriousness of their relationships as observed by the researcher. At the time of data collection, the participants' relationship status was what is indicated in Table 1. However, at the time of writing, all five couples were married and living together, and couples 1,2, and 5 had children together, which may be taken as an indication of their commitment to their relationships.
(3) NS SI was born in California. His parents were born in La Paz, El Salvador, and they raised their son in a Spanish-speaking household.
(4) NSE3 was born in California. When he was two years old, his family moved to Minnesota, where he was raised.
(5) The participants' self-reported amount of time spent speaking Spanish with their partners must be understood as a nonscientific estimate. However, it seems safe to assume that speakers can accurately measure if they speak more in one language or equally in both. There is reason to doubt the precision of the self-reported numbers, but it seems unlikely to doubt the participants' ability to remember which language(s) they generally use and to what extent they use them.
(6) An additional factor that arises from using heterosexual partners in a study of this nature has to do with the participants' target. While the closeness of the romantic relationship implies that the NSEs would select the NSSs' dialect as their intonation target, the NSSs were not of the same sex as the NSEs. The ideal target, therefore, might be a speaker of the same dialect as the NSSs and the same sex as the NSEs.
(7) Although alpha is usually set at 0.05, it was judged best to set it higher for this stud}'. Raising the alpha level is done to avoid a Type 2 error, which was considered more serious than a Type 1 error in this particular study. Committing a Type 2 error would indicate a similarity between the two speakers when, in reality, none actually exists. A more detailed explanation of Type 1 and 2 errors in t-tests can be found in Utts & Heckard (2006).
(8) Not all of the labels used are specific to the Sp_ToBI system (e.g. Aguilar et al. 2009). However, due to the contrast observed between the Spanish and English initial pitch accents and the initial and medial pitch accents, additional notifications were required to indicate the change in valley alignment.
Table 1. NSE/NSS personal information. NSE Sex Birthplace Time ACTFL studying proficiency Spanish 1 Female Illinois 1.5 years Novice high 2 Female Iowa 11 years Advanced low 3 Male California (4) 14 years Advanced high 4 Female Minnesota 10 years Advanced high 5 Female Minnesota 17 years Advanced high NSS Relationship Sex Birthplace Status Duration 1 Male La Paz, El Salvador (3) Dating 2 years 2 Male Guayaquil, Ecuador Engaged 3 years 3 Female Merida, Mexico Married 5 years 4 Male Merida, Mexico Engaged 4 years 5 Male Cuautla, Mexico Married 3 years Table 2. Language use in NSE/NSS daily interactions. What language do you On an average day, I speak speak with your partner? in Spanish--% of the time with my partner NSE1 Mostly English 1% NSE2 Spanish and English equally 40% NSE3 Spanish and English equally 45% NSE4 Mostly Spanish 90% NSE5 Spanish and English equally 50% On an average day, I speak in Spanish--hours and-- minutes with my partner NSE1 30 minutes NSE2 2 hours and 30 minutes NSE3 3 hours and 30 minutes NSE4 5 hours NSE5 1 hour Table 3. Total number of declaratives produced by speakers. Speaker Language Total declaratives NSE1 Spanish 22 English 29 NSS1 Spanish 30 NSE2 Spanish 28 English 30 NSS2 Spanish 28 NSE3 Spanish 29 English 29 NSS3 Spanish 30 NSE4 Spanish 30 English 30 NSS4 Spanish 28 NSE5 Spanish 29 English 28 NSS5 Spanish 23 Table 4. Percentage of deaccenting per stressed syllable. Speaker Language Deaccenting Initial Medial Final NSE1 Spanish 0% 14% 14% English 3% 35% 86% NSS1 Spanish 0% 0% 97% NSE2 Spanish 0% 0% 0% English 7% 63% 97% NSS2 Spanish 0% 14% 4% NSE3 Spanish 0% 14% 3% English 0% 34% 14% NSS3 Spanish 0% 27% 93% NSE4 Spanish 0% 33% 100% English 0% 20% 100% NSS4 Spanish 0% 75% 93% NSE5 Spanish 0% 12% 31% English 0% 0% 100% NSS5 Spanish 4% 73% 8% Table 5. Initial/medial stressed syllable scaling change. Speaker Language Peak scaling P-value Valley P-value change scaling change Mean Std. Mean Std. dev. dev. NSE1 Spanish 0.51 0.24 0.11 0.15 English -0.06 0.84 0.159 0.01 0.17 0.150 NSS1 Spanish 0.66 0.24 0.262 0.19 0.14 0.131 NSE2 Spanish 0.37 0.16 -0.02 0.09 English 0.05 0.66 0.250 0.20 0.19 0.034 * NSS2 Spanish 0.40 0.16 0.453 0.04 0.13 0.201 NSE3 Spanish 0.35 0.28 0.05 0.08 English 0.50 0.25 0.056 * 0.16 0.16 0.156 NSS3 Spanish 1.26 0.36 0.001 * 0.00 0.13 0.364 NSE4 Spanish 0.87 0.16 0.05 0.11 English 0.69 0.12 0.011 * -0.05 0.12 0.142 NSS4 Spanish 0.41 0.05 0.008 * 0.05 0.10 0.484 NSE5 Spanish 0.48 0.30 0.11 0.30 English 0.45 0.48 0.993 0.18 0.19 0.464 NSS5 Spanish 0.33 0.35 0.139 0.01 0.28 0.209 Table 6. Initial stressed syllable pitch accent alignment. Speaker Language Peak p-value Valley p-value alignment alignment Mean Std. Mean Std. dev. dev. NSE1 Spanish 123.5 98.0 -5.5 89.7 English 120.7 44.9 0.930 86.8 47.7 0.028 * NSS1 Spanish 141.9 19.4 0.533 -7.7 58.3 0.909 NSE2 Spanish 137.9 25.4 -22.3 37.4 English 153.7 60.4 0.387 142.2 78.4 <0.001 * NSS2 Spanish 72.3 46.5 0.001 * 7.64 18.7 0.029 * NSE3 Spanish 3.6 29.5 -41.4 50.0 English 47.6 67.5 0.088 * 40.3 46.2 0.003 * NSS3 Spanish 124.8 35.5 <0.001 * -79.1 30.1 0.021 * NSE4 Spanish 116.3 25.5 -58.5 74.0 English 176.5 37.7 0.003 * 34.9 75.6 0.002 * NSS4 Spanish 93.9 30.4 0.124 -37.4 37.7 0.356 NSE5 Spanish 86.0 22.2 -15.5 51.3 English 201.0 39.7 <0.001 * 91.4 71.7 <0.001 * NSS5 Spanish 45.0 16.5 <0.001 * -68.7 29.5 0.006 * Table 7. Medial stressed syllable pitch accent alignment. Speaker Language Peak p-value Valley p-value alignment alignment Mean Std. Mean Std. dev. dev. NSE1 Spanish 27.9 82.4 4.1 68.5 English 75.4 71.0 0.257 71.2 72.1 0.096 * NSS1 Spanish 125.6 34.3 0.007 * 55.6 43.7 0.041 * NSE2 Spanish 108.4 14.1 -6.4 39.7 English 150.4 58.3 0.072 * 148.1 80.7 0.007 * NSS2 Spanish 7.4 32.4 <0.001 * 17.3 15.5 0.200 NSE3 Spanish 64.5 35.7 3.09 27.5 English 25.9 56.7 0.074 * 31.5 44.3 0.141 NSS3 Spanish 100.6 75.1 0.248 28.8 42.2 0.024 * NSE4 Spanish 77.3 24.1 97.2 38.8 English 242.3 5.9 <0.001 * 117.7 27.4 0.127 NSS4 Spanish 77.3 37.7 0.356 -1.8 22.3 0.013 * NSE5 Spanish 64.7 32.4 67.3 44.5 English 247.7 27.7 <0.001 * 143.6 35.5 <0.001 * NSS5 Spanish -3.4 47.6 0.028 * -31.6 29.1 0.032 * Table 8. Pitch accent types produced by speakers. Speaker Language Pitch accent type Initial Medial Final NSE1 Spanish L+>H* !L*+>!H L+!H* English L*+>H L*+>H N/A NSS1 Spanish L+>H* !L*+>!H N/A NSE2 Spanish L+>H* L+>!H* L+!H* English L*+>H !L*+>H N/A NSS2 Spanish L*+>H L*+>!H L*+!H NSE3 Spanish L+>H* !L*+>!H L+!H* English L*+>H !L*+>!H L*+H NSS3 Spanish L+>H* L*+>!H N/A NSE4 Spanish L+>H* !L*+>!H N/A English L*+>H L*+>!H N/A NSS4 Spanish L+>H* !L+>!H* N/A NSE5 Spanish L+>H* !L*+>!H !L*+!H English L*+>H !L*+>!H N/A NSS5 Spanish L+>H* L+!H* !L*+H Table 9. Similarity of NSE Spanish to control groups. NSE1 NSE2 NSE3 NSE4 NSE5 Overall English X X [check] [check] X deaccenting Spanish X X X X X Peak scaling English [check] [check] X X [check] change Spanish [check] [check] X X [check] Valley scaling English [check] X [check] [check] [check] change Spanish [check] [check] [check] [check] [check] Initial peak English [check] [check] X X X alignment Spanish [check] X X [check] X Initial valley English X X X X X alignment Spanish [check] X X [check] X Medial peak English [check] X X X X alignment Spanish X X [check] [check] X Medial valley English X X [check] [check] X alignment Spanish X [check] X X X
|Printer friendly Cite/link Email Feedback|
|Author:||Peace, Meghann M.|
|Publication:||International journal of the Linguistic Association of the Southwest|
|Date:||Jun 1, 2013|
|Previous Article:||CALL research and practice: quo vadis?|
|Next Article:||Code-switching in social network messages: a case study of a bilingual Chicana.|