Next Article in Journal
A Study of Grammatical Gradience in Relation to the Distributional Properties of Verbal Nouns in Scottish Gaelic
Previous Article in Journal
Very Young Children Learning German Notice the Incorrect Syllable Stress of Words
Previous Article in Special Issue
Second Language Learner Attitudes Towards Peer Use of a Variable Sociophonetic Cue
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Listening for Region: Phonetic Cue Sensitivity and Sociolinguistic Development in L2 Spanish

by
Lauren B. Schmidt
Department of Spanish and Portuguese, San Diego State University, San Diego, CA 92182, USA
Languages 2025, 10(8), 198; https://doi.org/10.3390/languages10080198
Submission received: 8 April 2025 / Revised: 24 July 2025 / Accepted: 4 August 2025 / Published: 20 August 2025
(This article belongs to the Special Issue Second Language Acquisition and Sociolinguistic Studies)

Abstract

This study investigates how second language (L2) learners of Spanish identify the regional origin of native Spanish speakers and whether specific phonetic cues predict dialect identification accuracy across proficiency levels. Situated within a growing body of work on sociolinguistic competence, this research addresses the development of learners’ ability to use linguistic forms not only for communication but also for social interpretation. A dialect identification task was administered to 111 American English-speaking learners of Spanish and 19 native Spanish speakers. Participants heard sentence-length stimuli targeting regional phonetic features and selected the speaker’s country of origin. While L2 learners were able to identify regional dialects above chance, accuracy was low and significantly below that of native speakers. Higher-proficiency learners demonstrated improved identification, especially for speakers from Spain and Argentina, and relied more on salient phonetic cues (e.g., [θ], [ʃ]). No significant development was found for identification of Mexican or Puerto Rican varieties. Unlike native speakers, L2 learners did not show sensitivity to broader macrodialect groupings; instead, they frequently defaulted to high-exposure varieties (e.g., Spain, Mexico) regardless of the phonetic cues present. Findings suggest that sociophonetic perception in L2 Spanish develops gradually and unevenly, shaped by cue salience and exposure.

1. Introduction

The ability to participate fully across contexts of interaction in a second language (L2) entails a wide range of abilities that may be difficult to acquire through classroom input and interaction alone (Canale & Swain, 1980; Tarone & Swain, 1995). In addition to knowledge of grammatical rules and the ability to produce sounds in a native-like manner, native speakers can interpret information about the speaker, the context of interaction, and so on, often at a subconscious level (e.g., Campbell-Kibler, 2010). This extralinguistic information influences comprehension and future production of the language itself (Casasanto, 2008; Squires, 2014). In short, social facts about language users are an important part of linguistic competence for language learners and native speakers alike. For example, a Spanish speaker’s realization of /s/, as a sibilant or a reduced or elided variant, or the manner in which a French-speaker expresses verbal negation, with the presence or absence of the particle ne (‘not’), can provide information about the speaker’s country of origin, age, gender, social class and the formality of the context of interaction (Alba, 2004; Coveney, 1996; File-Muriel, 2009). Failure to acquire these abilities can result in misinterpreting social cues in the discourse setting and an inability to interact in socially, as well as linguistically, appropriate ways (Bardovi-Harlig, 2001).
A speaker’s real-time perception of variants in the moment of interaction becomes intricately linked to their individual world views and attitudes over the course of a lifetime (Drager, 2010). Perceiving a regional accent activates attitudes toward individuals from that region, which may extend beyond the interaction itself (Garrett et al., 2003). For example, hearers may form beliefs about an individual speaker’s personality, intrinsic abilities, and social status based solely on the way that speaker uses language. It has even been shown that social attitudes may go so far as to shape perception of a particular variant (Drager, 2005; Hay & Drager, 2010; Niedzielski, 1999). Notably, the ability to perceive these differences across speakers and interactions does not always imply the ability to articulate these facts, as speakers may not be consciously aware of the judgments made during real-time language processing (Mack & Munson, 2012). In fact, native speakers are generally quite bad at identifying dialects that are not their own, especially if they have resided in the same location all their lives (C. G. Clopper & Pisoni, 2004; Williams et al., 1999). Even so, reactions to (but not necessarily identification of) regional varieties are an expected component of nativelike competence. Moreover, better dialect recognition benefits L2 learners by enhancing intercultural sensitivity, deepening understanding of regional identities, and improving speech processing and listening comprehension in real-world contexts. For example, accurately recognizing the speaker’s dialect can enable listeners to better anticipate phonetic variation (specific to that dialect), thereby reducing processing load and improving understanding.
In second language learning, however, most existing research has examined learners’ acquisition of grammar and phonology for effective communication (Kissling, 2015; Lord, 2010). Recently, we have also seen growth in research on the acquisition of sociolinguistic variation in various target languages, and this extends to knowledge and use of variants that are constrained by social and linguistic variables (Dewaele, 2004; Hansen Edwards, 2020; Li, 2010; Mougeon et al., 2010; Solon & Kanwit, 2025), as well as the perception and production of regional variants (e.g., the interdental fricative in Spain, (George, 2014; Ringer-Hilfinger, 2012) or intervocalic /d/ elision in Andalusian Spanish, (Regan, 2022)). What remains less clear is how second language learners come to perceive and use specific linguistic features to interpret social information about a speaker and/or a context—such as a speaker’s region of origin—and whether this ability develops systematically with proficiency and experience.

2. Dialect Identification

2.1. Regional Variation and Dialect Recognition

There is growing evidence that social and linguistic information is stored together and used simultaneously during real-time language processing and production (e.g., C. G. Clopper & Pisoni, 2004). This is consistent with experience-based models of language, such as exemplar theory, which holds that information about the manner and context in which a form was used, the elements that co-occurred with that form and the linguistic patterns found in language are processed and stored together (e.g., Bybee, 2006). The geographic origin of the speaker is one example of such extralinguistic information. In cases of geographic variation, speakers from the same region share patterns of use of particular forms, such as the interdental fricative /θ/ in central Spain. Taking the example of a form that varies along geographic and social dimensions, speakers who aspirate syllable-final /s/ in Spanish (e.g., patas [ˈpa.tah] ‘paws’) can be identified as hailing from particular geographic regions, including southern Spain and the Caribbean, among many others, and we further expect greater rates of use among men, those with lower levels of education and in informal settings (e.g., Alba, 2004; File-Muriel, 2009). These examples demonstrate the complexity of the information conveyed through language and also the confluence of factors that interact as meaning is expressed and comprehended.
Speakers of a language are generally good at identifying their own dialectal group. However, outside one’s own regional variety, the ability to match a speaker to their region of origin diminishes. For example, C. G. Clopper and Pisoni (2004) showed that participants in Indiana matched American English speakers to region of origin with a relatively low (25%), but above chance (17%), accuracy rate, and Díaz-Campos and Navarro-Galisteo (2009) found that Spaniards and Venezuelans were similarly inaccurate in identifying Spanish speakers’ region of origin (34% average accuracy), but also above chance (17%). Dialect identification is mediated by several factors. For example, speakers with different native languages tend to rely on different cues when identifying different varieties, and certain regions tend to be evaluated more accurately because they are found to diverge more from others or to be associated with social meaning (e.g., Gooskens & van Bezooijen, 1999). This variation in accuracy may be linked to the salience of specific phonetic features, especially those that are socially meaningful or marked. Sumner et al. (2014) refer to differences in socioacoustic salience, explaining that some phones capture greater attention because of the social value they carry. For example, a socially marked pronunciation (like [ʃ] for [ʝ] in Argentinian Spanish) might stand out more strongly to listeners and influence judgments on speaker identity. It is also the case that characteristics of the individual (listener), such as language experience, mediate one’s ability to identify region of origin. C. G. Clopper and Pisoni’s (2004) so-called “army brats”, who had moved at least three times during their lives, were better at matching speakers with region of origin. Likewise, early exposure to different geographic varieties had a lasting effect.

2.2. Second Language Dialect Identification

For L2 learners, the use of dialectal features in L2 production depends on factors beyond mere exposure, which is necessary but insufficient for their incorporation and use (Goldstein, 1987). For example, age of arrival, degree of contact with native speakers, sensitivity to the standard variety, noticing the variants, affective responses to the features, and learner agency in production all play a role in whether a learner will produce nativelike variation (Friesner & Dinkin, 2006; Goldstein, 1987). In perception, extensive research on L2 English dialect recognition (e.g., Carrie & McKenzie, 2018; Clark & Schleef, 2010; Eisenstein, 1982; Georgiou, 2024; Jarvella et al., 2001; Ladegaard, 1998; McKenzie, 2008; Scales et al., 2006; Stephan, 1997) shows that learners can discriminate between L2 dialects, but are generally better at discrimination than identification (e.g., Cunningham-Andersson, 1996; Scales et al., 2006), particularly in free identification tasks (e.g., Clark & Schleef, 2010). Moreover, learners’ identification accuracy varies by dialect (e.g., Ladegaard, 1998), likely due to listener familiarity and the cultural prominence of certain varieties globally (e.g., Clark & Schleef, 2010; Ladegaard, 1998; McKenzie, 2008).
L2 identification of dialects of target languages other than English, however, is much less studied, with the exceptions of Lam and O’Brien’s (2014) and Kaiser et al.’s (2019) analyses of L2 German, Cunningham-Andersson’s (1996) study of L2 Swedish, and Schoonmaker-Gates’ (2018) study of L2 Spanish. Like the L2 English learners, L2 learners of these other target languages were able to discriminate between dialects to some extent but, in general, displayed low dialect recognition accuracies. In the case of L2 Spanish, Schoonmaker-Gates (2018) found that the learners enrolled in three different levels of undergraduate Spanish courses (2nd semester, 3rd or 4th semester, and at least 5th semester) were better at identifying speakers from Spain and Argentina than from all other regions tested (Colombia and Venezuela, Cuba, Mexico, and Peru), and that learner dialect familiarity resulted in more accurate dialect identifications.
Some individual differences in dialect identification correlate with learner proficiency and length of stay in a target language region (e.g., Cunningham-Andersson, 1996; Eisenstein, 1982). Nevertheless, learners need more than to study abroad to develop L2 dialectal competence. Schoonmaker-Gates (2017) found that both exposure to and explicit instruction of dialectal features aid classroom learners in attending to cues necessary for comprehension and the perceptual development of dialect. Moreover, although L2 learners are able to identify dialects, it appears that learners may attend to different cues than native speakers do and that their social evaluations of different dialects also differ (e.g., Clark & Schleef, 2010; Eisenstein, 1986; Kaiser et al., 2019; Ladegaard, 1998; McKenzie, 2008; Wirtz & Ender, 2025), although these social evaluations may more closely approach those of native speakers as proficiency increases (e.g., Eisenstein, 1986; Schmidt & Geeslin, 2022). The current study expands on this previous work to explore what specific phonetic cues in the speech signal L2 learners are using to make dialect identifications in the target language and how this develops over time.

2.3. Current Study and Research Questions

The current study aims to answer the following research questions (RQs). The first RQ is a replication of Schoonmaker-Gates’ (2018) previous work for a learner population in a public Midwestern university. The second RQ tests L2 socio-indexing of specific phonetic cues for speaker region of origin, and the development of these cues across L2 level.
RQ1. 
Can second language learners of Spanish correctly identify the region of origin of a Spanish speaker, and does identification ability improve with increased language proficiency?
RQ2. 
Does the presence of specific phonetic features predict accuracy in L2 dialect identification, and, if so, how does proficiency in the L2 interact with the use of these phonetic cues?

3. Materials and Methods

A dialect identification task and a verbal guise task, grammar task, and background questionnaire were administered to L2 learners of Spanish and a native speaker group. Participants completed the following tasks (in the following order) as an in-class activity in their language classes, using headsets at individual computer stations at their home institution in the United States (L2 group) or in Spain (NS group): (a) verbal guise task, in which they rated speakers on either kindness or prestige characteristics, (b) dialect identification task, (c) language background questionnaire, and (d) multiple-choice grammar task. All tasks were presented and completed via Qualtrics (2015). The dialect identification and written tasks are described below. See Schmidt and Geeslin (2022) for further details regarding the verbal guise task, which is not part of the current study.

3.1. Experimental Tasks

3.1.1. Dialect Identification Task

The dialect identification assessed to what degree L2 learners could recognize a Spanish speaker’s region of origin and correctly label it, and how L2 proficiency level (or degree of experience with the target language) and phonetic cues present in the utterance might play a role in L2 dialect recognition. Participants heard sentence-length stimuli spoken by Spanish speakers from four regional dialects—two male, university-educated speakers from each region—and were asked to identify the speaker’s region of origin from a list of 20 Spanish-speaking regions, which included the 18 Spanish-speaking Latin American countries, the commonwealth of Puerto Rico, and Spain. All speech stimuli were produced by speakers of the same gender to help ensure that listener responses reflect dialectal features rather than reactions to speaker gender, which can influence perception through social expectations (e.g., C. G. Clopper & Pisoni, 2007; Strand, 1999). Male speakers were chosen as they are more likely to produce socially salient nonstandard variants, which tend to be less frequent in female speech (Labov, 1990). The four regional dialects targeted in the task were Castilian, Mexican, Caribbean and Rioplatense Spanish; these were chosen as they represent distinct macrodialects of the Spanish-speaking world (Henríquez Ureña, 1921), and thus L2 listeners may vary in their abilities to recognize each variety.
Three sentence-long stimuli (see Table 1) were designed to target specific Spanish dialectal sounds representative of the four macrodialects: Mexican Spanish (Merida, Mexico), Caribbean Spanish (San Juan, Puerto Rico), Rioplatense Spanish (Buenos Aires, Argentina), and Castilian Spanish (Valladolid, Spain).
Only one regional phonetic feature was present in each spoken stimulus, and the feature was repeated twice within the sentence. These features may index a particular region (or multiple regions, so called “macroregions”, discussed further below) and/or serve to distinguish one region from another. Stimuli were recorded in each speaker’s home dialectal region using a Tascam DR-05 recorder with a Shure W20 headset mic (44.1 Hz). Speech rate and volume were normalized across the stimuli. Stimuli were not randomized in the task, but an attempt was made to distribute the dialectal voices in the presentation of stimuli to avoid the presentation of the same dialect back-to-back.
Table 2 summarizes the segmental features targeted in the task. In the Caribbean macrodialect (Puerto Rico), stimuli included lenited-/s/, velarized-/n/, and lambdacism (lateralization of /ɾ/), which are common in Caribbean Spanish but also found in some dialects of Andalusia and coastal Latin America (Lipski, 1994); these features also show social and stylistic variation. Rioplatense (Argentina) stimuli included aspirated-/s/ and the assibilated palatal [ʃ], the latter feature geographically limited to Argentina and Uruguay (Lipski, 1994). Castilian (Spain) stimuli featured apical-/s/ and the interdental fricative [θ], both unique to Spain (Hualde, 2005). The remaining features—laminal-/s/, seseo (absence of [θ]), and yeísmo ([ʝ] for ‘y’ and ‘ll’)—are common across many geographic varieties (Lipski, 1994) and served as control features to the other region-specific variants.

3.1.2. Language Background Questionnaire and Grammar Task

Participants completed a background questionnaire and a 24-item multiple-choice grammar test. The questionnaire gathered details about their academic history, biographic details, languages they speak, previous Spanish language education, and exposure to different Spanish dialects. The grammar test assessed their understanding of commonly taught formal grammatical properties of Spanish. Since language proficiency can vary among students in the same classroom, this measure was included to verify that there were differences in proficiency across the L2 learners included in the study and was used to group learners into distinct proficiency groups (discussed further below).

3.2. Participants

Of the 121 American English-speaking L2 Spanish learners recruited, 10 were excluded for reporting a native language other than English. Additionally, 24 native speakers (NSs) of Spanish, university students in Seville, southern Spain, completed the same tasks. The NS group was included, first, to test the validity of the instruments used and, second, to provide just one of many possible NS baselines in recognition of the region of origin of the speakers in the current task. This NS group was selected because the researcher had access to the speech community. Data from 5 NSs, those who were not L1 speakers of Spanish or had not grown up in Spain, were not included in the analyses, with a final total NS group of 19.
All L2 participants were also university students, enrolled in classes in the Midwestern United States. To study the development of the recognition and identification of different regional varieties of Spanish, L2 learners were selected from four enrollment levels: 1st Year Spanish, 2nd Year Spanish, 3rd Year Introduction to Spanish Linguistics, and 4th Year topics courses in Spanish Literature and Linguistics. Results from the Grammar Task showed that 1st and 2nd Year Spanish students had similar scores, while students in 3rd and 4th Year Spanish topics courses scored higher than the lower groups, but not significantly differently from each other (one-way ANOVA, F(3, 265) = 26.470, p < 0.001). Consequently, the participants were regrouped based on their test scores, creating three groups: those who scored below the 25th percentile on the 24-item multiple-choice grammar task were placed in Group 1, those between the 25th and 75th percentiles in Group 2, and those above the 75th percentile in Group 3. A strong positive correlation was found between the original Enrollment Level (1st Year, 2nd Year, 3rd Year, and 4th Year classes) and L2 Proficiency Groupings (Level 1, Level 2, Level 3), r (111) = 0.592, p < 0.001. Background data for the three L2 proficiency groups and the NS group are presented in Table 3.
Nearly all learners in the highest proficiency group (Level 3) were Spanish majors or minors, while this was the case for about half of the learners in the mid-level group (Level 2) and for almost none in the lowest group (Level 1). The primary learning environment for these learners was classroom-based, though around one-third, mostly from the higher proficiency groups, had study abroad experiences. Most studied abroad in Spain (n = 17), while others reported experiences in Latin American countries (e.g., Costa Rica, Argentina, Nicaragua, Chile, and Puerto Rico). The average duration of these study abroad experiences was 3 months but ranged from 2 weeks to 1 year. Besides studying abroad, learners also reported dialect contact through family members or close contacts from different Spanish-speaking regions and additional travel to Spanish-speaking areas. The Seville NS group reported dialect contact via travel to the region or, more commonly, social contacts who spoke those varieties. Each participant was coded (in a categorical fashion) as having dialect contact (1) or not (0) for each one of the four varieties tested. Contact was established if the participant reported at least one of the dialect contact experiences—social contacts, study abroad, or travel to the region. Table 4 provides a summary of the percentage of individuals who reported dialect contact experiences for each of the four dialects included in the current study, by participant group.
Overall, L2 learners in this program were most likely to be exposed to Spanish speakers from Spain and Mexico. This is probably due to the higher number of instructors from these regions and more opportunities for study abroad in Spain or for travel to Mexico. Consequently, learners were not equally exposed to all Spanish varieties through teaching materials and instructor input, but this too likely represents large, public research universities in the Midwest. Several individuals in the NS group from southern Spain reported contact with Argentine but not Mexican or Puerto Rican speakers.

3.3. Analysis

In the task, each participant selected a region of origin (country1) for three sentences per speaker, totaling 24 judgments per participant. Dialect identification accuracy scores were then calculated for each participant for each of the four regional dialects by dividing the number of times origin was correctly identified for each dialect by six, the total number of times the dialect was heard (three sentences × two speakers of each dialect). This country-level analysis follows prior studies (e.g., Díaz-Campos & Navarro-Galisteo, 2009; Schoonmaker-Gates, 2018) to allow for comparison. Moreover, it reflects how L2 learners are often exposed to dialects in pedagogical settings.
In addition to the by-country analysis of identification accuracy, participants’ Spanish-speaking country response selections were also grouped into seven larger macrodialects (that is, larger geographic regions that share common linguistic features), modified from Henríquez Ureña’s (1921) macrodialect groupings. These groupings reflect patterns of shared linguistic features and historical–geographic relationships commonly recognized in dialectological research (e.g., Henríquez Ureña, 1921; Lipski, 1994). Moreover, this approach aligns with prior findings from perceptual dialectology, which show that listeners tend to cluster similar varieties based on shared phonetic features and perceived social proximity (e.g., C. G. Clopper & Pisoni, 2007; Williams et al., 1999). The macrodialect groupings included the following: (1) Mexico (Mexico), (2) Central America (Costa Rica, El Salvador, Guatemala, Honduras, Nicaragua, Panama), (3) the Caribbean (Cuba, Puerto Rico, the Dominican Republic, Venezuela), (4) Andean Spanish (Bolivia, Colombia, Ecuador, Peru), (5) Chile (Chile), (6) Rio de la Plata (Argentina, Uruguay, Paraguay), and (7) Spain (Spain). Similarly, macrodialect identification accuracy scores were calculated for each participant for each of the four dialects by dividing the number of times any related country within the macrodialect grouping was selected (coded as accurate responses) by six, the total number of times the macrodialect was heard (three sentences × two speakers of each macrodialect)2.
Linear mixed model analyses were conducted, first for country level (dialect identification accuracy) and then for macrodialect level (macrodialect identification accuracy), with listener group, speaker dialect, and targeted phonetic feature as fixed effects, and participant as a random effect.

4. Results

4.1. Dialect Identification

The L2 (n = 111) and NS (n = 19) dialect identification accuracy scores are presented in Table 5 by country and by macrodialect. The NS group identified speakers’ country of origin with 44.1% accuracy (SD = 49.7), compared to 14.9% for the L2 group (SD = 35.7). Given a 5% chance baseline (1 out of 20), both groups appear to be using details in the speech signal to assign regional origin above chance level. A linear mixed model revealed a significant interaction between listener group and speaker dialect (F(3, 2984) = 72.174, p < 0.001), as well as significant main effects of listener group (F(1, 128) = 178.325, p < 0.001) and of speaker dialect (F(3, 2984) = 139.452, p < 0.001). NS listeners were most accurate (near ceiling) in identifying speakers within their own country of origin (Spain), followed by identification of Argentine and then of Mexican speakers, and least accurate in identifying Puerto Rican guises, with significant differences in identification between each dialect group (all pairwise comparisons p < 0.001). The L2 group as whole showed a somewhat different pattern, with the most accurate identification of country of origin of the Spain and Mexican speakers (with no significant difference between the two dialects, p = 0.721), followed by lower identification accuracy of the Argentine speakers and lastly of the Puerto Rican speakers (p < 0.001). Regarding across-group comparisons, the NSs were significantly more accurate than the L2 group in all dialect identifications (at p < 0.001 for both Argentina and Spain comparisons, and p = 0.014 for Mexico), with the exception of identification of the Puerto Rican guises (p = 0.353 for the NS-L2 comparison), where both groups scored low and close to chance levels.
When considering accuracy in selecting a country within the same macrodialect region as the speaker, the accuracy rates were higher—particularly for the native group—with the NS group accurate 56.4% of the time (SD = 49.6) and the L2 group at 19.6% (SD = 39.7). (Note: chance level for macrodialect identification is approximately 14%, or one out of seven). A second linear mixed model was run with dialect identification accuracy for macrodialect, which also revealed significant main effects of listener group (F(1, 128) = 255.613, p < 0.001)) and of speaker dialect (F(3, 2984) = 49.555, p < 0.001) and a significant interaction between listener group and speaker dialect (F(3, 2984) = 39.082, p < 0.001). When considering macrodialect identification accuracy, we see a difference for the native listeners, who are most accurate in identifying the macrodialect Spain, followed by the Rioplatense and the Caribbean regions (which were not significantly different from one another, p = 0.288), and least accurate in identifying Mexico (with all other pairwise comparisons significant at p < 0.001). That is, when considering macrodialect groupings, it was revealed that the native listeners had higher identification accuracy of the Caribbean variety as opposed to specifically identifying the country, Puerto Rico. The L2 learners as a whole, however, showed no differences in identification accuracies across the macrodialects, with all macrodialect comparisons not significant (p = 0.123 Spain-Caribbean, p = 0.530 Spain-Rioplatense, and p = 1.000 all other comparisons).
In other words, for the native listeners in particular, speakers in the task are often identified as originating from other countries that share similar regional linguistic features (i.e., within the same macrodialect). For example, the NSs labeled the Argentine (Rioplatense) guises as Uruguayan at times (8% of the time) and often mislabeled the Puerto Rican (Caribbean) guises as Cuban (33% of the time). No such patterns of (mis)labeling within the same macrodialect region were observed for the L2 group; rather, L2 listeners tended to erroneously assign speakers from all four dialects as from Mexico or Spain. This may reflect a response bias, with the L2 listeners defaulting region of origin associations to those speaker groups with which they are most familiar and have had the most contact. Analyses below, thus, focus on country-level identification.

4.2. L2 Development of Dialect Identification

Next, we consider L2 development in region dialect identification. Figure 1 shows mean dialect identification scores according to speaker dialect across the three L2 proficiency groups. A linear mixed model revealed significant main effects of speaker dialect (F(3, 2543) = 30.911, p < 0.001) and of L2 level (F(2, 108) = 11.273, p < 0.001), and a significant interaction between the two (F(6, 2543) = 4.464, p < 0.001). For the across-group analysis, improvement in identification of speaker region of origin is observed as learners increase in exposure to and proficiency in the target language. First, for the overall identification accuracy differences across levels, it was found that L2 learners in Level 1 and in Level 2 performed similarly in overall region of origin (country) accuracy (p = 0.735), with mean overall identification of all dialects combined at 12% (SD = 32.4) and 14% accuracy (SD = 34.4), respectively. Level 3 learners scored significantly higher in dialect identification than Levels 1 and 2 (p < 0.001 for both pairwise comparisons), with a mean overall identification accuracy for Level 3 learners of 21% (SD = 40.6).
Second, regarding the interaction between speaker dialect and L2 level, it was revealed that the improvements in dialect identification at the highest L2 level were limited to the identification of the Argentine and Spain guises only (with pairwise comparisons between Level 3 and the lower levels for these dialects at p < 0.001). The Level 3 group accurately identified the Argentine guises 25% of the time (SD = 43.3), compared to 7% (SD = 26.2) and 9% (SD = 43.3) for Levels 1 and 2, respectively. These most advanced learners furthermore accurately identified the Spain guises 34% of the time (SD = 47.5), while learners at Levels 1 and 2 accurately identified speakers from Spain only 17% (SD = 37.3) and 21% (SD = 40.4) of the time, respectively. There were no significant changes in identification accuracy of the Mexican (M1 = 19%, SD1 = 39.4; M2 = 20%, SD2 = 40.0; M3 = 19%, SD3 = 39.0) or Puerto Rican guises (M1 = 4%, SD1 = 20.4; M2 = 6%, SD2 = 22.9; M3 = 5%, SD3 = 22.5) across L2 level (all pairwise comparisons p = 1.00).

4.3. Phonetic Cue Use in Spanish Dialect Identification

Next, we consider how targeted phonetic cues influence L2 dialect identification accuracy. Figure 2 and Figure 3 present the dialect identification accuracies for the L2 and NS listener groups, respectively, according to speaker dialect and sentence stimulus (i.e., targeted phonetic cue). Two linear mixed models were run, one for each listener group3, with dialect identification accuracy for country as the dependent variable, speaker dialect and sentence stimulus (phonetic cue) as fixed effects, and participant as a random effect.
The L2 model found significant main effects of speaker dialect (F(3, 2542) = 38.460, p < 0.001) and of sentence stimulus (targeted phonetic cue) (F(2, 2542) = 46.038, p < 0.001) and a significant interaction between dialect and sentence (F(6, 2542) = 47.954, p < 0.001). Post hoc analyses of the accuracy of the different sentence stimuli (phonetic cues) within each dialect showed several significant comparisons; the significant pair-wise comparison results are discussed as follows. First, L2 learners identified Argentine speakers more accurately in the sentence with the assibilated palatal [ʃ] cue (POLLO, M = 24.8%, SD = 43.3) than in either of the other two sentences (POSTRE with aspirated-s, M = 5.9%, SD = 23.5; PELICULA with seseo (M = 5.9%, SD = 23.5; p < 0.001 for both comparisons)). Second, learners were more accurate in identifying speakers from Spain with the sentence with the interdental fricative [θ] cue (PELICULA, M = 54.5%, SD = 49.9) than for either of the other two sentences (POSTRE with maintained apical-s, M = 4.1%, SD = 19.8; POLLO with yeísmo (M = 9.5%, SD = 29.3; p < 0.001 for both comparisons)).
Additionally, the L2 group was also significantly better in identifying speakers from Mexico when hearing the sentence stimulus with seseo (that is, absence of [θ]) (PELICULA, M = 24.8%, SD = 43.3) than when hearing the sentence stimulus with syllable-final maintained laminal-s (POSTRE, M = 16.7%, SD = 37.4; p = 0.026) or the sentence with yeísmo (that is, absence of assibilated palatal [ʃ]) (POLLO, M = 17.6%, SD = 38.1; approaching significance at p = 0.059). There were no significant effects of phonetic cue for L2 identification of the Puerto Rican guises; all sentence stimuli—with lenited-s (POSTRE, M = 5.4%, SD = 22.7), velarized-n (PELICULA, M = 5.9%, SD = 23.5), and lambdacism (POLLO, M = 4.5%, SD = 20.8)—were identified at chance levels.
The NSs model likewise showed significant main effects of speaker dialect (F(3, 426) = 113.689, p < 0.001) and of sentence stimulus (targeted phonetic cue) (F(2, 426) = 6.460, p = 0.002) and a significant interaction between dialect and sentence (F(6, 426) = 5.578, p < 0.001). These NS listeners from southern Spain, like the L2 group, were significantly more accurate in identifying speakers from Argentina for the sentence stimulus with the [ʃ] acoustic cue (POLLO, M = 76.3%, SD = 43.1) than for either of the other two stimuli (POSTRE, M = 36.8%, SD = 48.9; PELICULA, M = 23.7%, SD = 43.1; p < 0.001 for both comparisons). However, there were no other significant differences across sentence stimuli (that is, across targeted phonetic segmental cues) for the NS listeners in categorizing each of the other dialects included in the task.

4.4. L2 Development in Phonetic Cue Use

Finally, we examine L2 development in the use of segmental cues for region of origin judgments; see Table 6, which presents dialect identification accuracy according to speaker dialect, sentence stimulus (phonetic cue present), and L2 learner proficiency level.
Separate linear mixed models, with dialect identification accuracy as the dependent variable, sentence stimulus (phonetic feature) and L2 level as fixed effects, and participant as a random effect, were run for each of the four dialects targeted in the task to determine differences across L2 level in dialect identification in the presence of a particular phonetic cue.
First, neither the sentence stimulus (F(2, 1438) = 0.549, p = 0.866) nor L2 level (F(2, 108) = 0.189, p = 0.828) were found to have significant effects in the model for L2 dialect identification of the Puerto Rican guises. That is, L2 listeners—at all levels—did not vary in identification accuracies of the Puerto Rican speakers according to the presence of those different phonetic features manipulated in the guises (i.e., lenited-/s/, velarized-/n/, and lambdacism). This was also the case for the Mexican guises, with no significant effects of sentence stimulus (F(2, 657) = 1.781, p = 0.169) nor L2 level (F(2, 657) = 0.113, p = 0.893) on L2 dialect identifications of the Mexican speakers. However, significant interactions between sentence stimulus and L2 group in dialect identification accuracies were found in the models run for the Argentine (F(4, 549) = 18.384, p < 0.001) and the Spain guises (F(4, 548) = 10.944, p < 0.001), discussed as follows.
In identification of the Argentine guises, Level 1 learners showed no significant differences in dialect identification accuracy according to any of the phonetic cues present. However, learners at both Level 2 and Level 3 were more accurate in Argentine identifications with the assibilated palatal cue (POLLO) than for the sentences with aspirated-s (POSTRE, p = 0.009 Level 2, and p < 0.001 Level 3) or with seseo (PELICULA, p = 0.068 approaching significance Level 2, and p < 0.001 Level 3). For the Spain guises, L2 learners at each of the three levels, however, were significantly more accurate in Spain identifications with the /θ/ cue (PELICULA) over the apical-s (POSTRE) and yeísmo stimuli (POLLO) (all pairwise comparison p < 0.001).

5. Discussion

5.1. L2 Dialect Identification in Spanish

RQ1 examined whether L2 Spanish learners can identify speakers’ regions of origin and whether identification accuracy increases with proficiency. Findings echoed previous research (e.g., Carrie & McKenzie, 2018; Clark & Schleef, 2010; C. Clopper & Bradlow, 2009; Lam & O’Brien, 2014; Scales et al., 2006; Schoonmaker-Gates, 2018; Stephan, 1997): learners showed some discrimination ability but performed poorly overall in regional identification (with only 15% overall accuracy) and were not as accurate as native speaker listeners. Learners were most accurate in identification of the regional origin of the speakers from Spain and Mexico, followed by Argentina, and were unable to identify the region of origin of the Puerto Rican speakers (at chance levels). This finding is especially interesting in light of the lower ratings for prestige afforded by L2 leaners to this variety (Schmidt & Geeslin, 2022), underscoring the likelihood that these types of socio-evaluative ratings among learner groups may reflect additional constructs such as intelligibility and comprehensibility, mitigated by familiarity with the variety through classroom activities and social contacts.
Schoonmaker-Gates (2018) likewise found that second language learners of Spanish were most accurate in recognition of dialect speakers from Spain, particularly for those who had prior exposure to Peninsular Spanish, followed by identifying Argentine Spanish, and that learners were worst at identifying region of origin for Colombian and Caribbean speakers. The higher accuracy in the Spain and Mexico identifications observed in the current study could be influenced in part by task response effects, that is, as these learners were more exposed to speakers from Mexico and Spain, they may have selected these two nation responses as a sort of ‘default’ choice when uncertain in their response. (This may also explain the unexpected result that learners more often identified the Mexican dialect for those stimuli produced with seseo; in the absence of the Spain interdental fricative cue, the learners selected the ‘other’ default nation, Mexico).
This learner familiarity identification bias is further evidenced when comparing the L2 and NSs identification responses at a macrodialectal level (what Díaz-Campos & Navarro-Galisteo, 2009 refer to as ‘dialect clustering’). For example, the NS, but not the L2 learners, often (mis)identified the Puerto Rican speakers as Cuban—a neighboring variety of Caribbean Spanish, which shares many of the same linguistic features, such as s-weakening and r-lambdacism. C. Clopper and Bradlow (2009) similarly found that native (as well as non-native, to some degree) listeners exhibited perceptual clusters of regional dialects of American English. The L2 learners in the current study, on the other hand, did not display this macrodialect (or ‘dialect clustering’) sensitivity but rather tended to most often miscategorize speakers as from Mexico or Spain, regardless of the linguistic features present in the speech stimuli. C. Clopper and Bradlow suggest that the lower accuracy and ‘noisier clustering solutions’ by non-native listeners result from both non-native attention to unreliable or inappropriate acoustic-phonetic cues and from a lack of signal-independent cultural knowledge surrounding how ‘constellations of cues’—rather than single features—define a dialect in the non-native language (C. Clopper & Bradlow, 2009, p. 450).
Regarding learner development in dialect identification abilities, the current study found improvement in identification of Spanish dialects across L2 level, with the Level 3 learners better at dialect recognition than the Level 1 and 2 learners, who patterned similarly. In addition to scoring highest on a grammaticality measure, the Level 3 group also differed from the lower-level groups in that the majority of Level 3 learners had prior study abroad experience in a Spanish-speaking region (and, most often in Spain), they reported the most contact with various dialects, and they reported particularly increased contact with speakers from Mexico, Argentina, and Spain. Eisenstein (1982) likewise found that higher proficiency L2 learners of English were more accurate in a (same-different) dialect discrimination task and that exposure to different dialects improved discrimination ability, and Cunningham-Andersson (1996) found that length of residence in Sweden correlated with better performance on a dialect identification task for L2 learners of Swedish. Furthermore, while L2 level (based on course enrollment level) did not predict learner dialect identification accuracy in Schoonmaker-Gates’ (2018) study of learners of Spanish, reported dialect familiarity was a significant predictor of dialect identification at all levels. Thus, it may be that increased opportunities for contact with different dialects of the target language or instruction on regional differences (in upper-level linguistics-focused courses, for example), rather than higher overall proficiency and greater grammatical knowledge of the L2, may be driving the development across L2 level in dialect identification observed in the current study.
Notably, accuracy gains among advanced learners were limited to certain dialects of the target language. The Level 3 learners were significantly better than the Level 1 and Level 2 learners in identification of the speakers from Spain and from Argentina. However, there were no differences in abilities to identify speakers from Mexico or Puerto Rico across the L2 levels studied. We turn now to the sociophonetic cues analysis to understand this interaction between L2 level (experience) and dialect on learner dialect identifications.

5.2. L2 Use of Sociphonetic Cues in Dialect Identification in Spanish

RQ2 investigated how specific phonetic cues and proficiency level influence L2 dialect identification accuracy. The L2 learners were more adept at recognizing the Argentine dialect in those sentence stimuli with the assibilated palatal [ʃ] cue—at Levels 2 and 3—and more adept at identifying speakers as from Spain in the stimuli with the interdental fricative [θ] cue—at all three levels. Thus, the improvements in L2 identification of some varieties—Argentina and Spain—but not others—Mexico and Puerto Rico—appears to be due to learners at the higher levels making greater use of specific individual phonetic cues—[ʃ] and [θ]—when making their dialect judgements, in contrast to native speakers who rely on ‘constellations of cues’ (C. Clopper & Bradlow, 2009) (as seen in the native listeners’ highly accurate identification of all Spain stimuli, even in the absence of [θ]).
These L2 dialect identification results align identically with the metalinguistic dialect awareness knowledge reported in a written survey completed by U.S. Midwestern L2 learners of Spanish, spanning from beginners to near-native L2 speakers (Schmidt, 2022). In the survey, learners demonstrated greatest explicit awareness of Peninsular Spanish features, particularly the interdental fricative [θ], followed by the Argentine assibilated palatal [ʃ], with far less reporting of other features from Latin American varieties, such as s-weaking or r-lambdacism, only seen at the more advanced levels.
Learners did not use other regional cues—including Castilian apical-s, Argentine aspirated-s, and Puerto Rican lenited-s, velarized-n, and lambdacism—in their regional identifications. Not all linguistic features carry equal social weight; their impact on social perception is mediated by perceptual salience and the clarity of their association with social categories (Barnes, 2015). Interestingly, Chappell and Kanwit (2022) found that the more advanced L2 learners in their Matched Guise study, and especially those who had taken a phonetics course, did link weakened-s to Caribbean region of origin. This discrepancy in findings could be due in part to the tasks themselves—in Chappell and Kanwit (2022), listeners heard speakers from only two regions (Mexico and Puerto Rico) and were given only four response options (Mexico, the U.S., the Caribbean, or somewhere else/I don’t know), while in the current study, listeners had twenty Spanish-speaking regions to select from while identifying speakers from four different dialects, increasing task difficulty. Furthermore, the highest-level learners in the current study reported the least exposure to Puerto Rican Spanish of all regions tested. Taking together findings from these two studies, it seems that while advanced L2 learners of Spanish may be capable of perceptually ‘attending to’ the s-weakening phonetic cue, sufficient familiarity or exposure (and/or explicit instruction) may also be necessary to subsequently acquire its association with Caribbean (and other s-weakening) dialects.
What is clear is that familiarity or exposure alone cannot fully explain L2 dialect identification abilities. In the case of Mexico, the L2 learners, despite reporting greatest contact with and exposure to Mexican Spanish (at all Levels), did not improve in recognizing speakers from Mexico. The NS listeners, likewise, exhibited the lowest identification accuracy for the Mexico macrodialect. These findings may be due to the lack of distinctiveness and uniqueness of phonetic segmental features of general or standard Mexican Spanish. That is, the features presented in the current study in the Mexican stimuli—seseo, yeísmo, and (maintained-) laminal-s—are common across many regional varieties of Spanish and not unique to Mexico, and thus, are not ‘linguistically marked’4 (Milroy & Milroy, 1999).
Listeners also likely use other phonetic information in the speech signal to make social categorizations of geographic region, in addition to the singular segmental features targeted in the current task—such as other sounds and, especially, prosodic information and other global speech characteristics like speech rate (C. Clopper & Bradlow, 2009; Dellwo et al., 2015). The NSs from southern Spain showed ceiling levels of accuracy in identifying the Spain guises for all of the sentence stimuli, and likely used multiple cues—often implicitly or unconsciously—when hearing speakers of dialects with which they are familiar. Future research should tease out to what degree L2 learners also implicitly acquire such sociolinguistic perception, and to what degree they rely on explicit, metalinguistic knowledge of variation in the target language. Moreover, as proposed by Gudmestad and Kanwit (2025), further study is needed to explore L2 development of perception and evaluation of social information linked to language beyond region of origin—such as gender, race, sexuality, socioeconomic status—and how learners use this variation to construct identity and social meaning.

6. Conclusions

This study contributes to our understanding of sociophonetic development in L2 Spanish by showing that learners can identify regional dialects at above-chance levels, though their accuracy remains low compared to native speakers. Dialect identification abilities improved with proficiency, but only for dialects such as Castilian and Rioplatense Spanish, and primarily when salient, ‘linguistically marked’ segmental cues like [θ] and [ʃ] were present. Learners did not show similar gains for Mexican or Puerto Rican varieties, suggesting that salience and exposure likely interact to play key roles in dialect recognition. Moreover, while native speakers showed sensitivity to macrodialect groupings, often misidentifying dialects within broader regional clusters (e.g., identifying Puerto Rican speech as Cuban), L2 learners did not. Instead, learners frequently defaulted to familiar or high-exposure varieties, such as Spain and Mexico, regardless of the phonetic cues present. This suggests that L2 listeners may rely less on constellations of cues or macrodialect associations and more on surface-level familiarity and stereotypical cues when making dialect judgments.
Dialect perception requires more than phonetic discrimination—it also involves social experience, cultural knowledge, and the ability to link phonetic cues to sociolinguistic meaning (C. Clopper & Bradlow, 2009). Even in first language acquisition, sociolinguistic competence—especially dialect awareness and social indexing—develops gradually, with significant advances not occurring until adolescence and adulthood (Dossey et al., 2020). L2 learners may be able to detect acoustic differences but lack the social framework to map these onto dialect categories. These results indicate that while L2 learners are sensitive to certain regional phonetic features, their ability to map these cues onto social categories like region of origin remains limited and unevenly developed. The findings reinforce the need for greater inclusion of dialectal variation in L2 instruction and highlight the importance of both linguistic input and sociolinguistic context in shaping the development of dialect awareness. Future research should further explore how learners integrate multiple cues and social experiences over time, and how sociophonetic perception interacts with identity construction and broader social evaluations in L2 acquisition.
The ability to recognize regional dialects carries implications for L2 learners across social and pedagogical domains. On the one hand, dialect awareness may enhance listening comprehension, facilitate more appropriate interaction across speech communities, and support the development of sociolinguistic competence. On the other, it may also lead learners to internalize native speaker language ideologies or reinforce linguistic hierarchies if not guided by critical awareness. While limited dialect recognition may shield learners from such biases in early stages, it may also result in missed cues or misinterpretation of speaker identity. These findings suggest that dialect recognition should be addressed in L2 instruction in ways that both broaden learners’ exposure and promote reflection on the social meanings of variation.

Funding

This research received no external funding.

Institutional Review Board Statement

The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Institutional Review Board of Indiana University (protocol code 1501532086; 1 February 2015).

Informed Consent Statement

Informed consent was obtained from all subjects involved in the study.

Data Availability Statement

The data presented in this study are available on request from the author due to privacy issues.

Acknowledgments

I would like to acknowledge the late Kimberly Geeslin, who was instrumental in the inspiration, design, and early analysis of this study. Her influence is deeply felt in this work.

Conflicts of Interest

The author declares no conflicts of interest.

Notes

1
The term country is used throughout for consistency and simplicity, including Puerto Rico, which, while an unincorporated U.S. territory and not a sovereign nation, is treated here as a distinct geopolitical entity due to its unique sociolinguistic profile.
2
Mexico and Spain each represent both an individual country and a macrodialect in this analysis. Unlike the Caribbean and Rioplatense groups, where a nearby variety could still be counted as correct at the macrodialect level, no such overlap applies to Mexico or Spain, which creates an asymmetry in scoring that may contribute to lower macrodialect identification rates for these varieties.
3
L2 proficiency and cue use were analyzed separately from NS to isolate developmental patterns. NS perceptual patterns are expected to vary across native speech communities.
4
Certainly, other types of linguistic features are unique and distinctive to Mexico, such as the use of the pronoun le for affective/emphatic purposes (e.g., échale ganas ‘give it your best’), and may similarly ‘linguistically mark’ Mexican Spanish.

References

  1. Alba, M. (2004). Fonética y fonología del español moderno: Guía didáctica. Universidad Autónoma de Santo Domingo. [Google Scholar]
  2. Bardovi-Harlig, K. (2001). Evaluating the empirical evidence: Grounds for instruction in pragmatics? In K. R. Rose, & G. Kasper (Eds.), Pragmatics in language teaching (pp. 13–32). Cambridge University Press. [Google Scholar] [CrossRef]
  3. Barnes, S. (2015). Perceptual salience and social categorization of contact features in Asturian Spanish. Studies in Hispanic and Lusophone Linguistics, 8(2), 213–241. [Google Scholar] [CrossRef]
  4. Bybee, J. (2006). From usage to grammar: The mind’s response to repetition. Language, 82(4), 711–733. [Google Scholar] [CrossRef]
  5. Campbell-Kibler, K. (2010). The sociolinguistic variant as a carrier of social meaning. Language Variation and Change, 22(3), 423–441. [Google Scholar] [CrossRef]
  6. Canale, M., & Swain, M. (1980). Theoretical bases of communicative approaches to second language teaching and testing. Applied Linguistics, 1(1), 1–47. [Google Scholar] [CrossRef]
  7. Carrie, E., & McKenzie, R. M. (2018). American or British? L2 speakers’ recognition and evaluations of accent features in English. Journal of Multilingual and Multicultural Development, 39(4), 293–307. [Google Scholar] [CrossRef]
  8. Casasanto, D. (2008). Embodiment of abstract concepts: Good and bad in right-and left-handers. Journal of Experimental Psychology: General, 137(3), 351–367. [Google Scholar] [CrossRef]
  9. Chappell, W., & Kanwit, M. (2022). Do learners connect sociophonetic variation with regional and social characteristics?: The case of L2 perception of Spanish aspiration. Studies in Second Language Acquisition, 44(1), 185–209. [Google Scholar] [CrossRef]
  10. Clark, L., & Schleef, E. (2010). The acquisition of sociolinguistic competence by German learners of English: A study of dialect comprehension and attitudes. Journal of Sociolinguistics, 14(5), 588–614. [Google Scholar] [CrossRef]
  11. Clopper, C., & Bradlow, A. (2009). Free classification of American English dialects by native and non-native listeners. Journal of Phonetics, 37, 436–451. [Google Scholar] [CrossRef]
  12. Clopper, C. G., & Pisoni, D. B. (2004). Homebodies and army brats: Some effects of early linguistic experience and residential history on dialect categorization. Language Variation and Change, 16(1), 31–48. [Google Scholar] [CrossRef]
  13. Clopper, C. G., & Pisoni, D. B. (2007). Free classification of regional dialects of American English. Journal of Phonetics, 35(3), 421–438. [Google Scholar] [CrossRef]
  14. Coveney, A. (1996). Variability in spoken French: A sociolinguistic study of interrogation and negation. Elm Bank Publications. [Google Scholar]
  15. Cunningham-Andersson, U. (1996). Learning to interpret sociodialectal cues. Speech, Music, and Hearing: Quarterly Progress and Status Report, 37(2), 155–158. [Google Scholar]
  16. Dellwo, V., Leemann, A., & Kolly, M.-J. (2015). The role of segments and prosody in the identification of a speaker’s regional origin. Journal of Phonetics, 48, 13–24. [Google Scholar] [CrossRef]
  17. Dewaele, J.-M. (2004). The acquisition of sociolinguistic competence in French as a foreign language: An overview. In F. Myles, & R. Towell (Eds.), The acquisition of French as a second language (pp. 120–139). John Benjamins. [Google Scholar]
  18. Díaz-Campos, M., & Navarro-Galisteo, I. (2009). Perceptual categorization of dialect variation in Spanish. In J. Collentine, M. García, B. Lafford, & F. Marcos Marín (Eds.), Selected proceedings of the 11th hispanic linguistics symposium (pp. 179–195). Cascadilla Proceedings Project. [Google Scholar]
  19. Dossey, E., Clopper, C. G., & Wagner, L. (2020). The development of sociolinguistic competence across the lifespan: Three domains of regional dialect perception. Language Learning and Development, 16(4), 330–350. [Google Scholar] [CrossRef]
  20. Drager, K. (2005). From bad to bed: The relationship between perceived merger and speech production. Te Reo, 48, 55–68. [Google Scholar]
  21. Drager, K. (2010). Sociophonetic variation in speech perception. Language and Linguistics Compass, 4(7), 473–480. [Google Scholar] [CrossRef]
  22. Eisenstein, M. (1982). A study of social variation in adult second language acquisition. Language Learning, 32(2), 231–252. [Google Scholar] [CrossRef]
  23. Eisenstein, M. (1986). Target language variation and second-language acquisition: Learning English in New York City. World Englishes, 5, 31–46. [Google Scholar] [CrossRef]
  24. File-Muriel, R. (2009). Acoustic cues of lenition in the Spanish of Barranquilla, Colombia. Hispania, 92(4), 850–872. [Google Scholar]
  25. Friesner, M., & Dinkin, A. J. (2006, May 27–30). The acquisition of sociolinguistic competence in a study abroad context: The case of L2 learners of French. 2006 Annual Conference of the Canadian Linguistic Association, North York, ON, Canada. [Google Scholar]
  26. Garrett, P., Coupland, N., & Williams, A. (2003). Investigating language attitudes: Social meanings of dialect, ethnicity and performance. University of Wales Press. [Google Scholar]
  27. George, A. (2014). Study abroad in central Spain: The development of regional phonological features. Foreign Language Annals, 47(1), 97–114. [Google Scholar] [CrossRef]
  28. Georgiou, G. P. (2024). Categorization of second-language accents by bilingual and multilingual listeners. British Journal of Developmental Psychology, 42(3), 425–438. [Google Scholar] [CrossRef]
  29. Goldstein, L. M. (1987). The role of language exposure in second language phonological acquisition [Doctoral Dissertation, Cornell University]. [Google Scholar]
  30. Gooskens, C. S., & van Bezooijen, R. (1999). Identification of language varieties: The contribution of different linguistic levels. Journal of Language and Social Psychology, 18(1), 31–48. [Google Scholar] [CrossRef]
  31. Gudmestad, A., & Kanwit, M. (2025). Reconsidering the social in language learning: A state of the science and an agenda for future research in variationist SLA. Languages, 10(4), 64. [Google Scholar] [CrossRef]
  32. Hansen Edwards, J. (2020). The acquisition of English L2 sociolinguistic variation. In D. Edmonds-Wathen, & S. Chelliah (Eds.), The routledge handbook of language revitalization (pp. 485–495). Routledge. [Google Scholar]
  33. Hay, J., & Drager, K. (2010). Stuffed toys and speech perception. Linguistics, 48(4), 865–892. [Google Scholar] [CrossRef]
  34. Henríquez Ureña, P. (1921). Observaciones sobre el español en América. Revista de Filología Española, 8, 357–390. [Google Scholar]
  35. Hualde, J. I. (2005). The sounds of Spanish. Cambridge University Press. [Google Scholar]
  36. Jarvella, R., Bang, E., Jakobsen, A. L., & Mees, I. M. (2001). Of mouths and men: Non-native listeners’ identification and evaluation of varieties of English. International Journal of Applied Linguistics, 11, 37–56. [Google Scholar] [CrossRef]
  37. Kaiser, I., Ender, A., & Kasberger, G. (2019). Varietäten des österreichischen Deutsch aus der HörerInnenperspektive: Diskriminations-fähigkeiten und sozio-indexikalische. In L. Bülow, A. Fischer, & K. Herbert (Eds.), Dimensionen des sprachlichen Raums: Variation—Mehrsprachigkeit—Konzeptualisierung (pp. 341–362). Peter Lang Verlag. Schriften zur deutschen Sprache in Österreich. [Google Scholar]
  38. Kissling, E. M. (2015). Teaching pronunciation: Is explicit phonetics instruction beneficial for FL learners? The Modern Language Journal, 99(3), 519–535. [Google Scholar] [CrossRef]
  39. Labov, W. (1990). The intersection of sex and social class in the course of linguistic change. Language Variation and Change, 2(2), 205–254. [Google Scholar] [CrossRef]
  40. Ladegaard, H. J. (1998). National stereotypes and language attitudes: The perception of British, American and Australian language and culture in Denmark. Language and Communication, 18(4), 251–274. [Google Scholar] [CrossRef]
  41. Lam, H., & O’Brien, M. G. (2014). Perceptual dialectology in second language learners of German. System, 46, 151–162. [Google Scholar] [CrossRef]
  42. Li, X. (2010). Sociolinguistic variation in the speech of learners of Chinese as a second language. Language Learning, 60(2), 366–408. [Google Scholar] [CrossRef]
  43. Lipski, J. M. (1994). Latin American Spanish. Longman. [Google Scholar]
  44. Lord, G. (2010). The combined effects of immersion and instruction on second language pronunciation. Foreign Language Annals, 43(3), 488–503. [Google Scholar] [CrossRef]
  45. Mack, S., & Munson, B. (2012). The influence of /s/ quality on ratings of men’s sexual orientation: Sounding gay and sounding straight. Journal of Phonetics, 40(1), 198–212. [Google Scholar] [CrossRef]
  46. McKenzie, R. M. (2008). The role of variety recognition in Japanese university students’ attitudes towards English speech varieties. Journal of Multilingual and Multicultural Development, 29(2), 139–153. [Google Scholar] [CrossRef]
  47. Milroy, J., & Milroy, L. (1999). Authority in language: Investigating standard English. Routledge. [Google Scholar]
  48. Mougeon, R., Nadasdi, T., & Rehner, K. (2010). The sociolinguistic competence of immersion students. Multilingual Matters. [Google Scholar]
  49. Niedzielski, N. (1999). The effect of social information on the perception of sociolinguistic variables. Journal of Language and Social Psychology, 18(1), 62–85. [Google Scholar] [CrossRef]
  50. Qualtrics. (2015). Qualtrics [Computer software]. Qualtrics. Available online: https://www.qualtrics.com (accessed on 20 July 2025).
  51. Regan, B. (2022). Individual differences in the acquisition of language-specific and dialect-specific allophones of intervocalic /d/ by L2 and heritage Spanish speakers studying abroad in Sevilla. Studies in Second Language Acquisition, 45(1), 65–92. [Google Scholar] [CrossRef]
  52. Ringer-Hilfinger, K. (2012). Learner acquisition of dialect variation in a study abroad context: The case of the Spanish [θ]. Foreign Language Annals, 45(3), 1–17. [Google Scholar] [CrossRef]
  53. Scales, J., Wennerstrom, A., Richard, D., & Wu, S. H. (2006). Language learners’ perceptions of accent. TESOL Quarterly, 40(4), 715–738. [Google Scholar] [CrossRef]
  54. Schmidt, L. B. (2022). L2 development of dialect awareness in Spanish. Hispania, 105(2), 267–284. [Google Scholar] [CrossRef]
  55. Schmidt, L. B., & Geeslin, K. L. (2022). Acquisition of linguistic competence: Development of sociolinguistic evaluations of regional varieties in a second language. Revista Española de Lingüística Aplicada, 35(1), 206–235. [Google Scholar] [CrossRef]
  56. Schoonmaker-Gates, E. (2017). Regional variation in the language classroom and beyond: Mapping learners’ developing dialectal competence. Foreign Language Annals, 50(1), 177–194. [Google Scholar] [CrossRef]
  57. Schoonmaker-Gates, E. (2018). Dialect comprehension and identification in L2 Spanish: Familiarity and type of exposure. Studies in Hispanic and Lusophone Linguistics, 11(1), 193–214. [Google Scholar] [CrossRef]
  58. Solon, M., & Kanwit, M. (2025). Exploring the relationship between preference and production as indicators of L2 sociophonetic competence. Languages, 10(4), 65. [Google Scholar] [CrossRef]
  59. Squires, L. (2014). Processing, evaluation, knowledge: Testing the perception of English subject–verb agreement variation. Journal of English Linguistics, 42(2), 144–172. [Google Scholar] [CrossRef]
  60. Stephan, C. (1997). The unknown Englishes? Testing German students’ ability to identify varieties of English. In E. W. Schneider (Ed.), Englishes around the world: Studies in honour of manfred görlach (pp. 93–108). John Benjamins Publishing. [Google Scholar]
  61. Strand, E. A. (1999). Uncovering the role of gender stereotypes in speech perception. Journal of Language and Social Psychology, 18(1), 86–100. [Google Scholar] [CrossRef]
  62. Sumner, M., Kim, S. K., King, E., & McGowan, K. B. (2014). The socially weighted encoding of spoken words: A dual-route approach to speech perception. Frontiers in Psychology, 4, 1015. [Google Scholar] [CrossRef]
  63. Tarone, E., & Swain, M. (1995). A sociolinguistic perspective on second language use in immersion classrooms. The Modern Language Journal, 79(2), 166–178. [Google Scholar] [CrossRef]
  64. Williams, A., Garrett, P., & Coupland, N. (1999). Dialect recognition. In D. Preston (Ed.), Handbook of perceptual dialectology (Vol. 1, pp. 345–358). John Benjamins. [Google Scholar]
  65. Wirtz, M. A., & Ender, A. (2025). Functional prestige in sociolinguistic evaluative judgements among adult second language speakers in Austria: Evidence from perception. Languages, 10(4), 67. [Google Scholar] [CrossRef]
Figure 1. Dialect identification accuracy according to speaker dialect and L2 level (country level).
Figure 1. Dialect identification accuracy according to speaker dialect and L2 level (country level).
Languages 10 00198 g001
Figure 2. L2 dialect identification accuracy according to phonetic cue/stimulus (country level).
Figure 2. L2 dialect identification accuracy according to phonetic cue/stimulus (country level).
Languages 10 00198 g002
Figure 3. NSs dialect identification accuracy according to phonetic cue/stimulus (country level).
Figure 3. NSs dialect identification accuracy according to phonetic cue/stimulus (country level).
Languages 10 00198 g003
Table 1. List of stimuli in dialect identification task.
Table 1. List of stimuli in dialect identification task.
# (No.)Stimulus
1 (POSTRE)Pablo pidió un postre de chocolate para la fiesta.
‘Pablo ordered a chocolate dessert for the party.’
2 (PELÍCULA)La nueva película de acción glorificó a la nación.
‘The new action film glorified the nation.’
3 (POLLO)Comí pollo y carne en la calle del norte.
‘I ate chicken and meat on north street.’
Table 2. Targeted phonetic features in task stimuli for each dialect group of speakers.
Table 2. Targeted phonetic features in task stimuli for each dialect group of speakers.
DialectMacrodialectPOSTREPELÍCULAPOLLO
Mexico aMexican(maintained-) laminal-/s/
postre [ˈpos-tɾe]
seseo
acción [ak-ˈsi̯on]
yeísmo
pollo [ˈpo-ʝo]
Puerto RicoCaribbeanlenited-/s/ b
postre [ˈpo-tɾe]
seseo; velar-/n/
acción [ak-ˈsi̯oŋ]
yeísmo; lambdacism
pollo [ˈpo-ʝo];
carne [ˈkal-ne]
ArgentinaRioplatenseaspirated-/s/
postre [ˈpoh-tɾe]
seseo
acción [ak-ˈsi̯on]
assibilated palatal
pollo [ˈpo-ʃo]
SpainCastilianapical-/s/
postre [ˈpos̪-tɾe]
interdental fricative
acción [ak-ˈθi̯on]
yeísmo
pollo [ˈpo-ʝo]
a Yucatecan Mexican Spanish is characterized by other region-specific features; however, these did not appear in the task stimuli. b Syllable-final pre-consonantal /s/ may be lenited to aspiration or deletion in Puerto Rican Spanish; an acoustic analysis of the task stimuli confirmed a majority of word-internal deleted-/s/ realizations for both speakers. One speaker also produced a single token of word-internal aspirated-/s/.
Table 3. Participants’ background data.
Table 3. Participants’ background data.
GroupnM Age (SD)Grammar Score M (SD, Range)% Span Majors (n), Minors (n)N Hours Span × Week% Study Abroad (n)
L2 Level 12719.7 (1.34)6.9 (1.27; 3–8)7% (0, 2)1 (1.35)4% (1)
L2 Level 25919.8 (1.39)11.2 (1.54; 9–14)49% (11, 18)2 (2.50)19% (11)
L2 Level 32520.0 (1.35)17.6 (2.49; 15–23)92% (14, 9)2 (2.43)68% (17)
NS (Spain)1922.8 (5.10)23.1 (1.23; 19–24)---
Table 4. Reported dialect contact according to participant group.
Table 4. Reported dialect contact according to participant group.
GroupnMexico ContactPuerto Rico ContactArgentina ContactSpain Contact
L2 Level 12730% (8)11% (3)4% (1)15% (4)
L2 Level 25949% (29)19% (11)9% (5)32% (19)
L2 Level 32568% (17)12% (3)24% (6)72% (18)
NS (Spain)195% (1)0% (0)37% (7)100% (19)
Table 5. Dialect identification accuracy scores (SD) according to speaker dialect and listener group.
Table 5. Dialect identification accuracy scores (SD) according to speaker dialect and listener group.
Dialect (Country/Macrodialect)L2 Accuracy Country 1L2 Accuracy Macrodialect 2NS Accuracy Country 1NS Accuracy Macrodialect 2
Mexico/Mexico19.7% (39.6)19.7% (39.6)28.9% (45.6)28.9% (45.6)
Puerto Rico/Caribbean5.3% (22.2)17.5% (38.0)8.8% (28.4)46.5% (50.1)
Argentina/Rioplatense12.2% (32.5)18.8% (39.1)45.6% (50.0)57.0% (49.7)
Spain/Spain22.7% (41.9)22.7% (41.9)93.0% (25.7)93.0% (25.7)
Total14.9% (35.6)19.6% (39.7)44.1% (49.7)56.4% (49.6)
1 Chance in accurate identification of country of origin is 5%. 2 Chance in accurate identification of macrodialect of origin is 14%.
Table 6. Dialect identification accuracy scores (SD) across L2 level according to speaker dialect and sentence stimulus.
Table 6. Dialect identification accuracy scores (SD) across L2 level according to speaker dialect and sentence stimulus.
DialectSentenceTargeted FeatureLevel 1Level 2Level 3
MexicoPOSTRElaminal-s22% (42.0)16% (36.9)12% (32.8)
PELÍCULAseseo22% (42.0)27% (44.6)22% (41.8)
POLLOyeísmo13% (33.9)18% (38.4)22% (41.8)
Puerto RicoPOSTRElenited-s0% (0.0)7% (25.2)8% (27.3)
PELÍCULAvelarized-n7% (26.3)6% (23.7)4% (19.8)
POLLOlambdacism6% (23.0)4% (19.2)4% (19.7)
ArgentinaPOSTREaspirated-s4% (19.0)4% (20.2)10% (30.4)
PELÍCULAseseo7% (26.3)7% (25.2)2% (14.1)
POLLOassib. palatal /ʃ/11% (31.6)15% (36.0)62% (48.8)
SpainPOSTREapical-s4% (19.0)3% (18.1)6% (24.0)
PELÍCULAinter. fric. /θ/43% (49.7)46% (49.9)88% (32.7)
POLLOyeísmo4% (19.0)13% (33.4)8% (27.3)
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Schmidt, L.B. Listening for Region: Phonetic Cue Sensitivity and Sociolinguistic Development in L2 Spanish. Languages 2025, 10, 198. https://doi.org/10.3390/languages10080198

AMA Style

Schmidt LB. Listening for Region: Phonetic Cue Sensitivity and Sociolinguistic Development in L2 Spanish. Languages. 2025; 10(8):198. https://doi.org/10.3390/languages10080198

Chicago/Turabian Style

Schmidt, Lauren B. 2025. "Listening for Region: Phonetic Cue Sensitivity and Sociolinguistic Development in L2 Spanish" Languages 10, no. 8: 198. https://doi.org/10.3390/languages10080198

APA Style

Schmidt, L. B. (2025). Listening for Region: Phonetic Cue Sensitivity and Sociolinguistic Development in L2 Spanish. Languages, 10(8), 198. https://doi.org/10.3390/languages10080198

Article Metrics

Back to TopTop