The Second Language Acquisition of Second-Person Singular Forms of Address: Navigating Usage and Perception in a Tripartite System in Medellin, Colombia

Abstract

Previous studies have found that second language learners can acquire sociolinguistic variation. However, there is a lack of studies that examine the L2 acquisition of second-person singular forms of address (2PS) in Spanish, especially in the immersion context of study abroad. The current study examines the acquisition of Spanish 2PS by seven adults learning Spanish in Medellin, Colombia. Participants completed an oral discourse completion task and a matched guise task to measure language perceptions toward each 2PS. Learners’ results are compared to findings from 38 native Spanish speakers from Medellin. Learners produced very few instances of the local variant vos and overproduced tú, differing greatly from native speakers. Two factors were found to significantly condition 2PS usage for learners: speaker gender and interlocutor relationship. Findings show that although learners perceive vos to a somewhat native-like extent and the role that it plays in the local variety, learners do not actually use it.

1. Introduction

The acquisition of Spanish second-person singular forms of address (2PS) is a complex process, especially for learners whose first language only has one 2PS, such as English. This is true for many reasons. First, the obvious reason is the fact that English only has one 2PS “you”. Spanish, in contrast, has either two or three depending on the dialect. Most Spanish 2PS systems include the informal “tú” and the formal “usted”, while some dialects also have “vos”. There are of course some dialects that use vos categorically, such as the case in the Rioplatense area of Argentina. Thus, learners of Spanish have to learn when and with whom to use each 2PS. Not taking vos into account and only looking at tú and usted, one might think that this would be rather easy to acquire. It is true that when age and power line up, it is rather simple. For example, an older teacher would be addressed with “usted”, or a classmate of the same age would be addressed with “tú”. In addition, the distinction between tú and usted varies regionally, which can be confusing for learners. However, as Dufon (2010) points out, it is a much more complex process to acquire native-like usage of 2PS in ambiguous situations, such as when power and age do not line up. This could be, for example, if your boss is younger than you or if an employee of yours is older. Does one use tú or usted? If vos is also an option, it becomes very complicated for a learner of Spanish whose first language is English. This is the case in the Spanish of Medellin, as the 2PS system is tripartite consisting of tú, vos, and usted. It should be noted that 2PS usage in Medellin is variable—these three forms exist, and how they are used is variable. There is a lack of research examining the L2 acquisition of a tripartite Spanish 2PS system, especially in the immersion context. In particular, tripartite 2PS systems provide a unique opportunity to examine the acquisition of the cultural nuances, the degree of understanding of the social dynamics between speakers, and the sociopragmatic competence required for the proper use of each form of address, while L2 learners are immersed in the target dialect. Therefore, a main goal of the current study is to investigate the L2 acquisition of Spanish 2PS with learners taking Spanish classes in Medellin.

In the field of sociolinguistics, until recently, many studies have tended to focus on production but not as much on perception. Examining perception in addition to usage provides a wider window into a speaker’s linguistic variety. Being able to examine a learner’s perception of a linguistic phenomenon can help inform us about their acquisition. For example, it may be that learners have not yet arrived at an advanced enough level to produce a linguistic phenomenon, but they might perceive it and understand its social value. Since most sociolinguistic and second language acquisition (SLA) studies do not investigate perception of specific variables, a second goal of the present study is to examine learners’ perceptions of Spanish 2PS and compare them to those of native Spanish speakers from Medellin in order to gain a more complete picture of these learners’ acquisition of Spanish 2PS. Based on the findings of the current study, we find that although learners perceive vos to a somewhat native-like extent and the role that it plays in the local variety, learners do not actually use it.

2. Review of the Literature

2.1. 2PS in Spanish

Unlike English, which only has one 2PS, “you”, Spanish and other Romance languages have multiple 2PS which are split by what Brown and Gilman (1960) refer to as the conflicting axes of power and solidarity. In their framework, T pronouns (e.g., tú in Spanish and tu in French) show familiarity or solidarity while V pronouns (e.g., usted in Spanish and vous in French) indicate politeness. In Spanish, tú and usted are present in virtually all Spanish dialects. Tú is the informal pronoun, whereas usted is considered the formal pronoun. In some dialects, vos is also an option. Vos is generally thought to be like tú in that it is very informal and generally shows great familiarity with the speaker. Voseo, the usage of the pronoun vos, and/or its verbal morphology, is found in several Spanish-speaking countries to varying degrees. In the same vein, the terms tuteo and ustedeo denote the usage of tu/usted and their verbal morphology. Regional voseo occurs in Peru to a limited extent among indigenous speakers, internal western Cuba, the Chiapas province of Mexico, Andean Venezuela, parts of southwestern and central Colombia, Ecuador, and Bolivia (Kapović, 2007; Lipski, 1994; Penny, 2000; Placencia, 1997). In contrast, a more generalized voseo is present in Argentina, Uruguay, Paraguay, and many parts of Central America (Penny, 2000).

2.2. 2PS in the Spanish of Medellin

The 2PS system in the Spanish of Medellin, Colombia has several distinct characteristics. One such characteristic is the usage of voseo, which results in a tripartite system of 2PS that consists of tú, usted, and vos. Examples of each 2PS and their corresponding verbal inflection, prepositional pronoun, and object pronoun are depicted in

Table 1. 2PS in Medellin, Colombia (from Newall, 2016).

Variant	Verbal Inflection	Prepositional Pronoun	Object Pronoun
tú	cantas ‘sing’	para ti ‘for you’	te veo ‘I see you’
vos	cantás ‘sing’	para vos ‘for you’	te veo ‘I see you’
usted	canta ‘sing’	para usted ‘for you’	lo/le/la veo ‘I see you’

.

As can be seen in Table 1, each 2PS has its own verbal inflection. However, it is important to note that in various instances, the verbal inflection for tú and vos is the same. This is the case for the past tenses and one-syllable present tense verbs, such as “vas” (‘you go’) or “estás” (‘you are’). In addition, it is important to mention that Spanish is a pro-drop language, and thus the subject pronouns (tú, vos, usted) are not obligatory. Therefore, it is possible and common for the verbal inflection to occur by itself without each subject pronoun being expressed explicitly.

Jang (2013) affirms that voseo is a strong symbol of identity for Paisas (people from the province of Antioquia, Colombia, of which Medellin is the capital). Jang (2013) posits that he does not expect voseo to disappear due to the strong sense of belonging that Paisas have and the negative attitudes that they have toward other Colombian dialects that do not employ voseo, such as the case in the capital city, Bogota. Voseo is a way that Paisas use to distinguish themselves from Colombians of other dialects (Jang, 2013).

Another interesting characteristic of the 2PS system in Medellin is that usted is used in both formal and informal situations, resulting in a dual usted. As mentioned above, generally in Spanish, usted is the formal 2PS. However, in the Spanish of Medellin, the use of usted/ustedeo has extended to be used not only with unknown or nonintimate interlocutors but also with intimate interlocutors, such as with spouses, one’s children, and even pets (Uber, 1985). In addition to the dual usted, polymorphism of 2PS, the usage of more than one 2PS directed toward the same person in the same interaction, is also common in Paisa Spanish (Denbaum-Restrepo, 2023a).

2.3. The Second Language Acquisition of 2PS

Dufon (2010) points out that acquiring the 2PS system is a complex process for learners. They first must acquire the 2PS system for unambiguous situations and then for ambiguous situations. For example, when age and power do not line up (e.g., a professor around the student’s age or a classmate that is much older than a learner), this creates an ambiguous situation. Dufon (2010) posits that acquiring 2PS in unambiguous situations is possible just from reading textbooks, but for the ambiguous situations, learners must engage in social interaction with native speakers. In addition, Dufon (2010) explains that acquisition of the 2PS system depends on the quantity and quality of learners’ relationships with these native speakers.

One of the first studies conducted that examined the L2 acquisition of forms of address was Warnick’s (1991) study, which examined L2 Japanese students. He found that students with higher proficiency used more appropriate first- and second-person forms of address. However, overall, the learners were considered non-native-like because they used more 2PS than native Japanese speakers and they often selected 2PS that were inappropriate for the context. Other studies have investigated 2PS in L2 French and L2 Spanish.

Lyster and Rebuffot (2002) and Dewaele (2004) both investigated the acquisition of French 2PS for second language (L2) learners. The context for Lyster and Rebuffot’s (2002) study was a French immersion class. They found very little use of vous by the learners and observed that the teacher used the tu form when referring to the entire class. Thus, the researchers concluded that the almost exclusive use of tu was due to the little input of vous that students received. Dewaele (2004) compared 2PS usage with native French speakers and learners. Results showed that 13 out of 52 learners used vous categorically, while 8/9 of the native speakers used tu exclusively. In addition, he found that learners sometimes switched between tu and vous in the same conversation, which native speakers did not do at all.

Other studies have examined the L2 Spanish acquisition of 2PS (Fernández, 2016; González-Lloret, 2008; Villarreal, 2014). Villarreal (2014) examined L2 Spanish production of 2PS in role plays and found an overgeneralization of tú. Similar to what Dewaele (2004) found with L2 French, Villarreal’s (2014) results showed an alternation of forms that could not be explained as strategic. Fernández (2016) examined L2 acquisition in the study abroad context. The participants were conversation partners, consisting of a Spanish L2 speaker and a native Argentinian, over four months. Data were collected from face-to-face interactions as well as text chat conversations. Before arriving in Argentina, the learner was aware of the use of vos, which is widely used in Argentina. However, her conversation partner used tú with her, and thus, the learner also used tú. Upon realizing that her conversation partner used tú with her because she was a Spanish teacher used to accommodating to students who are not aware of vos, both participants began to use vos in their interactions. González-Lloret (2008) also examined the development of 2PS with conversation partners of Spanish learners and native Spanish-speaking students from Spain, though in a different context: synchronous computer-mediated communication (SCMC). González-Lloret (2008) reports on the development of a case study consisting of two learners of Spanish communicating with a native Spanish-speaking student via Yahoo Messenger. The group of students worked together on a project over a period of 10 weeks. Results showed that the Spanish learners alternated between the formal usted and the informal tú, even though the native speaker explicitly told them not to call him “usted”. However, toward the end of the project, the main Spanish learners eventually started to adapt to the sociopragmatic norm established by the native speaker, using more tuteo and less ustedeo.

Overall, studies on L2 acquisition of 2PS show that learners in the immersion context generally lack native-like patterns of sociopragmatic competence. As stated previously, this is mainly seen by overgeneralizations of specific forms and alternations of various forms without justification. However, some research has shown more promising results. For instance, Pozzi (2021) found a strong association with the learners’ social network in the host country and the higher use of vos forms, particularly with high-proficiency students. Similarly, Fernández’s (2016) research in the study abroad context and González-Lloret’s (2008) research in the SCMC context also show a similar pattern, although it is important to recognize that the learner in Fernández’s (2016) study arrived in the host country already having a knowledge of the local form of address, vos. The acquisition of 2PS is important because it is both a sociolinguistic and sociopragmatic phenomenon. First, there are social dynamics between speakers that need to be understood on the part of the L2 learner so that proper address is used for a successful communicative act. In addition to the social dynamics, the acquisition of sociolinguistic conditionings in 2PS production is also important. We know that the L1 use of each form of address is constrained by social and linguistic factors, including speaker age, socioeconomic class, gender, subject pronoun expression, speech act, interlocutor relationship, and interlocutor age, to name a few (Newall, 2016; Denbaum-Restrepo & Restrepo-Ramos, 2022). In addition, the alternation of 2PS forms in the same discourse to address the same interlocutor has been found to be prevalent in tripartite dialects, such as Paisa Spanish (Denbaum-Restrepo, 2023a). All these factors involved in the production of tripartite 2PS systems present a compelling case to understand the level of sociopragmatic competence required by L2 learners in the acquisition of these forms.

2.4. Language Attitudes in Second Language Acquisition

The present study compares language attitudes and perceptions of native speakers from Medellin to those of learners taking Spanish classes in Medellin. Previous studies on language attitudes have been carried out with L1 and L2 speakers. For example, L1 studies have examined attitudes toward specific phonemes (Casillas, 2013; Díaz-Campos & Killam, 2012) or perceptions toward different dialects (Büdenbender, 2013). Casillas investigated the variation of the allophones /tʃ/ and /ʃ/, such as the pronunciation of “muchacho” as /mu.‘ʃa.ʃo/, in the Spanish in Tucson, Arizona. The researcher employed a matched guise that included audio clips of two male speakers and two female speakers reading from a script. There were two versions of audio clips for each speaker that were identical except for the use of the two allophones: one had /tʃ/, and the other had /ʃ/. Results showed that speakers who realized the fricative had lower ratings in terms of credibility, competence, and solidarity. Additionally, the listeners, undergraduate students in the Spanish program for heritage speakers, were able to distinguish which speakers were PhD students and which had only a high school education based on their respective usage of /tʃ/ and /ʃ/.

Díaz-Campos and Killam (2012) also examined language perceptions using a matched guise. They examined two specific phenomena: intervocalic /d/ deletion and syllable-final /ɾ/ deletion with participants from Caracas, Venezuela. The researchers found that the same speaker was evaluated differently based on whether they retained or elided syllable-final /ɾ/. Retention was evaluated significantly higher in terms of intelligence and professionalism, showing the prestige associated with the retention of syllable-final /ɾ/. In contrast, for intervocalic /d/, the same trends were observed regarding deletion and retention, but there were no significant differences. These results show the difference in social evaluations between the two variables. Findings from Díaz-Campos and Killam (2012) and Casillas (2013) demonstrate that native speakers have language attitudes associated with different phonetic variables.

In addition to native Spanish speakers having language attitudes toward specific linguistic variants, other studies have found that native speakers have varied perceptions toward specific dialects in general. For example, Büdenbender (2013) examined whether Puerto Ricans could identify Dominicans based purely on linguistic factors and how Dominicans were perceived in terms of education, social class, and pleasantness of speech. Büdenbender (2013) utilized a verbal guise in which seven Dominicans and one Puerto Rican were recorded responding to a question about how to prepare for a hurricane. Results showed that linguistic features played an important role in the identification of national origin. In addition, results suggested possible socioeconomic profiling in which Dominicans are attributed to having a lower socioeconomic class than Puerto Ricans.

Studies on L2 attitudes and perceptions have also been carried out, especially in the study abroad context. Both Knouse (2012) and Ringer-Hilfinger (2012) investigated the L2 acquisition of the interdental fricative /θ/ in study abroad contexts in Spain. In north-central Spain, the phonological phenomenon of distinción occurs, where the interdental fricative /θ/ is produced in words that include the graphemes <z>, <ce, ci> while the alveolar fricative /s/ is used for <s>. This is in contrast to Latin American dialects where /s/ is used for all four of these graphemes. Knouse (2012) measured language attitudes using a Pronunciation Attitude Inventory (PAI) in which students rated 12 statements about general pronunciation in an L2 based on a 5-point Likert scale. The more emphasis a student put on acquiring and employing native-like pronunciation, the higher the score was. It was expected that the higher the score on the PAI, the more likely it was that students would acquire /θ/. However, a higher PAI score did not correlate to increased production of /θ/. In addition, contact with native speakers failed to result in more /θ/ production. In contrast to Knouse’s (2012) study, for which language attitudes did not have a significant correlation to acquisition, Ringer-Hilfinger (2012) did find that language attitudes played a role in L2 acquisition of /θ/. Similar to Casillas (2013), Ringer-Hilfinger (2012) used a matched guise to measure language attitudes. The speakers for her matched guise consisted of six Spanish-English bilinguals: a female Colombian, a male Puerto Rican, a male Spaniard, a female Spaniard, a female learner, and a male learner. All speakers read the same text twice, once with distinción and once with seseo. As opposed to distinción, seseo refers to the use of /s/ in all contexts regardless of any grapheme distinction (e.g., <z>, <ce, ci>, <-d>. The idea was that participants thought they were listening to 12 different people. Participants listened to all 12 audio clips and rated them using a Likert scale from 1 to 6 on the following qualities: competence, integrity, and attractiveness. Overall results from the matched guise did not show a preference for distinción or seseo, and study abroad did not seem to influence learners’ attitudes toward use of /θ/. However, out of the two students who produced /θ/ on the production tasks, both had a positive attitude toward its use by NSs. Furthermore, there were learners who did not want to use /θ/ because they had more contact in the U.S. with Latinos and thought the /θ/ would be perceived negatively by them. It is possible that the differences in findings between Knouse (2012) and Ringer-Hilfinger (2012) could be due to the fact that the PAI was measuring L2 pronunciation in general and not specific to the host dialect, whereas the matched guise used by Ringer-Hilfinger (2012) did indeed measure perceptions toward a variant of the Castilian dialect—specifically, the same variant that she was examining in the production task.

In addition to examining L2 perceptions toward a specific variant, K. L. Geeslin and Schmidt (2018) investigated language attitudes toward various dialects in general. The researchers used a variation of the matched guise—a verbal guise. In this task, 110 Spanish learners listened to 24 sentence stimuli and rated the speakers on either kindness (“nice” and “kind”) or prestige (“intelligent” and “rich”) adjectives using a 6-point Likert scale. The stimuli for the verbal guise were recordings of eight male speakers, consisting of two speakers from each of the following dialects: Castilian Spanish, Rioplatense Spanish, Caribbean Spanish, and Mexican Spanish. Each speaker read the same three sentences, which were chosen carefully so that each sentence would exhibit the same number of dialectal cues for each dialect. Results showed that, overall, learners differentiated between dialects, as they rated them differently. Specifically, they rated Argentine Spanish as the most kind and Castilian Spanish as the least kind. In terms of prestige, Puerto Rican Spanish was rated as significantly lower than the other three dialects. The researchers then examined learners’ language attitudes broken down by study abroad location. Learners who studied in Spain showed similar ratings of kindness to the overall L2 group but higher prestige ratings for Castilian Spanish. Learners who studied abroad in Argentina had lower prestige levels for Argentinian Spanish but higher kindness levels, which the authors posit shows solidarity with Argentine speakers. One learner who studied abroad in Puerto Rico rated Puerto Rican Spanish as less kind, which was an unexpected result. The other learner who studied abroad in Puerto Rico rated the Puerto Rican dialect as more prestigious than the overall learner group. The authors explain that this is likely due to increased exposure to the Puerto Rican dialect, specifically exposure to its use in prestigious contexts. Another explanation is that it is likely that this learner developed greater comprehensibility of this dialect, which in turn could lead to higher prestige ratings.

Since previous studies have shown that both L1 and L2 speakers of Spanish distinguish various dialects based on linguistic phenomena and have language perceptions toward these dialects and specific linguistic variants, it seems worthwhile to further investigate the role of language perceptions in the L2 acquisition of Spanish 2PS. Additionally, K. L. Geeslin and Schmidt (2018) stress the “…importance of language attitudes in understanding learner orientation toward particular language-learning contexts and the value of investigating this dynamic more thoroughly as a key element of understanding the complex process of Second Language Acquisition”. The same authors expanded on these results in an article on L2 language attitudes, specifically on Peninsular Spanish varieties (Schmidt & Geeslin, 2022). Grammon (2021) also examines the complex nature of language attitudes and learner agency in acquiring dialect-specific/regional features. He found that the adoption of Peninsular Spanish [θ] was avoided by learners and the instructor to appeal against an idealized monolingual speaker. The author argues that language ideologies, along with sociolinguistic competence and national identity, affect L2 acquisition and production of dialectal variants.

The present study examines the L2 acquisition of 2PS among adults learning Spanish in an immersion context in Medellin, Colombia. It compares their usage to that of native Paisas, local speakers from Medellin, Colombia. In addition, it examines learners’ perceptions of Spanish 2PS and compares them to those of native speakers. Here, we compare our findings with results by Denbaum-Restrepo and Restrepo-Ramos (2022), who examined 2PS perceptions of native speakers from Medellin using the same tasks and methodology as the current study. Analyzing the 2PS perception of L2 learners gives us further insights into their developing grammar compared to monolingual speakers.

The present study is guided by the following research questions:

• What are the production frequencies of each 2PS (tú, vos, usted) for learners? How do learners’ rates compare to those of natives from Medellin?
• What linguistic and extralinguistic constraints condition the use of 2PS for learners? How do these linguistic constraints compare to native speakers?
• How do learners’ perceptions of each 2PS compare to native speakers’ perceptions?

3. Methodology

3.1. Participants

Participants consisted of a group of seven adult learners of Spanish who were taking Spanish classes in the Spanish as a Foreign Language Program at a university in Medellin, Colombia, and 38 native speakers of Paisa Spanish from Medellin. Learners were a heterogeneous group as their ages, Spanish levels, countries of origin, and time spent in Colombia varied.

Table 2. Summary of participants’ characteristics.

Participant	Age	Sex	Country of Origin	Spanish Level (CEFR)	Time Living in Medellin at Time of Data Collection	Total Time They Will Spend in Medellin
1	31	M	US	3 (A2)	2 months	1 year
2	25	M	US	9 (B2)	6 months	11 months
3	27	F	Germany	11 (B2)	10 months	1 year
4	24	F	US	9 (B2)	6 months	not sure
5	34	F	Canada	6 (B1)	1 year, 11 months	2.5 years
6	33	M	Bosnia	unknown	3 months	several years
7	39	M	Netherlands	11 (B2)	4 months	at least 2 years

summarizes these characteristics of participants.

As can be seen in Table 2, participants were from the following countries: Bosnia (N = 1), Canada (N = 1), Germany (N = 1), Netherlands (N = 1), and the US (N = 3). The variety of Spanish levels are based on the levels according to the Common European Framework of Reference (CEFR). Courses 1 and 2 correlate to level A1, and courses 3, 4, and 5 correspond to level A2. Courses 6, 7, and 8 correspond to level B1, while courses 9, 10, and 11 match up with level B2. Moreover, participants had spent a varied amount of time in Medellin at the time of data collection and planned on spending differing amounts of total time in Medellin. A trend that can be noted is that all participants were planning on spending a relatively long amount of time in Medellin altogether, as most participants were planning on staying in Medellin for at least 1 year. Furthermore, two learners planned on remaining in Medellin for 2 or 2 and a half years. Participants’ living situations were also heterogeneous. Some participants were living in an apartment with Colombian roommates (N = 3), while others lived in their own apartment by themselves (N = 2). No participants stayed with a Colombian host family, in a hotel or hostel, or in an apartment with other foreigners (non-Colombians). Some participants responded that they had a different living situation other than the ones listed above (N = 2).

The 38 native speakers were either born in Medellin and its surrounding municipalities or had moved as a child and, therefore, had lived in Medellin for the majority of their lives. Participants consisted of a stratified sample of education levels, genders, and ages. Native speakers’ ages ranged from 18–71 (average = 40.6), and 24 were female, while 18 were male. They came from a range of education levels including elementary (N = 5), high school (N = 16), and university or postgraduate education (N = 17). These speakers were part of a subset of participants from a 2022 study focusing on L1 sociolinguistic production and perception. They were selected for the present study based on the completion of the matched guise task.

3.2. Materials and Procedure

Oral Discourse Completion Task (DCT). For the DCT, participants were asked to read a context (accompanied by a photograph) and then respond (not narrate) with what they would say to the interlocutor in each situation (see Figure 1 for an example). The contexts manipulated the following four factors: speech act (complaint, command, question), relationship (intimate vs. distant), age of interlocutor (same age vs. older), and gender of interlocutor (female vs. male). This gave 24 possible combinations, and thus, the DCT consisted of 24 items.

Matched Guise. The matched guise was designed in order to measure (1) participants’ perceptions toward the three 2PS (vos, tú, and usted) by rating each speaker on how nice, honest, and professional he appeared and (2) participants’ perceptions of how close the relationship between the speaker and the interlocutor seemed.1 In a Qualtrics survey, participants heard a short sentence and were asked to rate the speaker using a response scale with five different options (see Figure 2).

Stimuli for the audios for the matched guise consisted of recordings from four speakers, each of whom were males from Medellin between the ages of 20 and 30. Each speaker read three sentences (each consisting of a different speech act: complaint, command, or question) three times each, varying the 2PS (and verbal form when applicable) as seen in

Table 3. Conditions and stimuli for the matched guise.

Speech Act	2PS	Stimulus
Reprimand	tú,	¿Por qué tú siempre haces eso? ‘Why do you (T) always do that?’
Reprimand	vos	¿Por qué vos siempre hacés eso? ‘Why do you (V) always do that?’
Reprimand	usted	¿Por qué usted siempre hace eso? ‘Why do you (U) always do that?’
Command	tú	Ven a la fiesta pues. ‘Come (T) to the party.’
Command	vos	Vení a la fiesta pues. ‘Come (V) to the party.’
Command	usted	Venga a la fiesta pues. ‘Come (U) to the party.’
Question	tú	¿Tú vives cerca de acá? ‘Do you (T) live close to here?’
Question	vos	¿Vos vivís cerca de acá? ‘Do you (V) live close to here?’
Question	usted	¿Usted vive cerca de acá? ‘Do you (U) live close to here?’

. These sentences were presented in isolation with no contextualization included. Audio clips were randomized, and 10 fillers were randomly inserted throughout. Fillers consisted of recordings of four female speakers between the ages of 18 and 22 from Spain and the Dominican Republic who read completely different sentences.2

Background Questionnaire. Participants also completed a background questionnaire in Qualtrics. This questionnaire asked participants general information including age, gender, place of birth, highest level of education, language background, Spanish level, other languages spoken, other countries lived in, living situation while in Medellin, time spent in Medellin, and total time anticipated to remain in Medellin.

3.3. Coding

3.3.1. Dependent Variable

The dependent variable of this study is the production of a 2PS (tú vs. usted vs. vos) or its verbal inflection (when unambiguous):

• Tú: “Hola papá, tengo que hacer un árbol genealógico, pero no sé todos nuestros antepasados. ¿Por eso puedes ayudarme por favor? Porque tú sabes toda la información y me gustaría hacerlo contigo”. ‘Hi Dad, I have to make a family tree, but I don’t know all of our relatives. For this reason, can you (T) help me please? Because you know (T) all the information and I would like to do it with you’ (Participant 3: German female, Spanish level 11)
• Usted: “Hola señor porfa ponga el botón por el piso tercero, gracias. O tercer piso”. ‘Hi sir, please put (U) the button for floor third, please. Or the third floor.’ (Participant 4: American female, Spanish level 9)
• Vos: ¿Por qué no venís a nuestra reunión?” ‘Why don’t you come (V) to our meeting?’ (Participant 7: Dutch male, Spanish level 11)

When the verbal morphology was ambiguous, these tokens were excluded. For example, (1) and (2) were excluded since it is ambiguous whether they are tuteo or voseo.

(1)

“¿Podrías decirme tu nombre otra vez por favor?” ‘Could you (T/V) tell me your name again please?’

(2)

“¿Estás aquí por la novia o el novio?” ‘Are you (T/V) here for the bride or the groom?’

3.3.2. Independent Variables

This study examines five independent variables. Descriptions of each variable follow.

The following variables were chosen when creating the DCT and were manipulated in a balanced way throughout. They were selected since previous research on native Spanish speakers has shown that they significantly constrain 2PS variation (e.g., Newall, 2016; Denbaum-Restrepo & Restrepo-Ramos, 2022).

• Speech act:

  ○

  Command: “Hey chica, chica cuídate. ¿Estás borracha o qué?” ‘Hey girl, girl take care of yourself (T). Are you drunk or what? (Participant 2: American male, Spanish level 9)

  ○

  Indirect command: “Hey, chica yo sé que tú estás nueva acá, pero somos un equipo y la próximo vez por favor me pasas la pelota para que podamos jugar todos juntos”. ‘Hey, girl I know that you are new here, but we are a team and next time please pass (T) me the ball so that we can all play together.’ (Participant 3: German female, Spanish level 11).

  ○

  Question: “¿Tú conoces la novia o el novio?” ‘Do you know (T) the bride or the groom?’ (Participant 4: American female, Spanish level 9)

  ○

  Declarative: “Mi amigo, necesitas repagarme y yo sé que tienes el dinero porque estás bailando en un discoteca muy cara y comiendo sushi entonces es muy obvio que tienes dinero para repagarme”. ‘My friend, you need (T) to pay me again and I know you have (T) the money because you are dancing in a very expensive club and eating sushi. So, it is very obvious that you have (T) money to pay me again.’ (Participant 4: American female, Spanish level 9)

• Relationship with interlocutor:

  ○

  Intimate: Le presté $400 dólares a mi amigo el mes pasado para que pudiera pagar la renta…Le digo a mi amigo… ‘I lent my friend $400 dollars last month so that he could pay the rent…I say to my friend…’

  ○

  Distant: Estoy en una fiesta con todos mis amigos. Nos encanta bailar, pero el dj, quien es de mi misma edad más o menos, está poniendo música muy mala…Cuando alcanzo el dj, le digo… ‘I am in a party with all my friends. We love to dance, but the DJ, who is around my same age, is playing really bad music…When I reach the DJ, I say to him…’

• Age of interlocutor:

  ○

  Same age: Estoy en clase de biología. Todo el mundo está hablando de sus cumpleaños y qué piensan hacer para sus fiestas. Yo estoy con mi mejor amigo y le cuento de lo que a mí me gustaría hacer para mi cumpleaños…Le pregunto a mi amigo… ‘I am in biology class. Everyone is talking about their birthdays and what they are planning on doing for their parties. I am with my best friend and I tell him what I would like to do for my birthday…I ask my best friend…’

  ○

  Older: Hay una señora mayor que está ayudando a entrenarme y enseñarme todo lo que tengo que saber…Le pregunto a la señora… ‘There is an older lady who is helping to train me and teach me everything that I have to know…I ask the lady…’

• Gender of interlocutor:

  ○

  Male: Tengo un compañero de cuarto que es un poco antisociable…Le digo a mi compañero de cuarto… ‘I have a (male) roommate that is a little antisocial…I told my (male) roommate…’

  ○

  Female: Hay un perro muy lindo y empiezo a hablar con su dueña, una chica de mi misma edad…Le pregunto a la chica… ‘There is a very cute dog and I start to talk to its (female) owner, a girl around my same age…I ask the girl…’

Additionally, data were coded for the following extralinguistic variable in

Table 4. Extralinguistic variables coded.

Variable	Variants
Speaker gender	Male
	Female

.

4. Results

4.1. Production Results

4.1.1. Overall Distribution

First, the overall distribution of 2PS from DCT responses by learners is presented below in

Table 5. Rates for 2PS production from the DCT by L2 learners.

2PS	Number	Percentage
Tuteo	106	75.2%
Ustedeo	32	22.7%
Voseo	3	2.1%
Total	141	100%

.

From Table 5, it can be observed that tuteo was produced the most and to a great extent. Ustedeo was produced the second most, while voseo was rarely used. For the sake of comparison,

Table 6. Rates of 2PS for native speakers (from Denbaum-Restrepo and Restrepo-Ramos (2022)).

2PS	Number	Percentage
Tuteo	148	40.5%
Ustedeo	158	37.4%
Voseo	149	22.0%
Total	452	100%

shows 2P3S rates for native speakers from Medellin from Denbaum-Restrepo and Restrepo-Ramos’ (2022) data.

Comparing the results for learners to those of native speakers, it can be observed that the learners share the same hierarchy for rates of 2PS. However, there are great differences. Mainly, native speakers have much more balanced usage between all three 2PS, whereas learners rely heavily on tuteo with some ustedeo and barely any voseo.

For the learners, there were only three voseo tokens altogether, which were produced by two participants. The three voseo tokens are presented below:

(3)

“Amiga, ¿por qué le contás al profesor que yo engañé? Tú sabías que no tenía tiempo para estudiar. Entonces eso es muy mal para mí y mis estudios”. ‘Friend, why do you tell (V) the professor that I cheated? You knew (T) that I didn’t have time to study. So that is very bad for me and my studies.’ (Participant 4: American female, Spanish level 9)

(4)

“Hola chica, estos son mis amigos y no quiero que perderlos entonces porfa. Entonces, porfa mudá a otra parte de la gente”. ‘Hi girl, these are my friends and I don’t want to lose them, so please. So, please move (V) to another part of the crowd.’ (Participant 4: American female, Spanish level 9)

(5)

“¿Por qué no venís a nuestra reunión?” ‘Why don’t you come (V) to our meeting?’ (Participant 7: Dutch male, Spanish level 11)

As can be seen in examples (3)–(5), all voseo tokens were only verbal morphology; no tokens of the explicit pronoun “vos” were produced by these learners.

4.1.2. Linguistic and Extralinguistic Constraints for Production of 2PS

To examine which linguistic and extralinguistic variables constrain production of 2PS in learners of Spanish in Medellin, a multivariate mixed-effects regression was carried out in Rbrul (Johnson, 2009). We used Rbrul because the program allows us to examine the effect of independent variables and include random effects, while displaying the results in terms of factor weights, a popular and easy-to-read output in studies of language variation. Since there were only three tokens of voseo, vos was excluded from the analysis and the dependent variable of 2PS produced only had two levels, consisting of either tuteo or ustedeo. Speech act, interlocutor gender, interlocutor relationship, interlocutor age, and speaker gender were run as fixed effects, while participant and DCT item were included as random effects. The application value was ustedeo.

Table 7. Multivariate mixed effects regression: ustedeo vs. tuteo.

Input	0.226
AIC	141.143
Total Tokens	137
	Factor Weight	Percentage	Number
Speaker Gender	p < 0.001
Female	0.67	32.9%	60
Male	0.34	11.9%	67
Interlocutor Relationship	p = 0.027
Distant	0.65	31.4%	70
Intimate	0.35	13.4%	67
Interlocutor Age	p = 0.080 *
Older	[0.62]	27.8%	79
Same Age	[0.38]	15.5%	58
Interlocutor Gender	p = 0.296 *
Male	[0.57]	27.3%	5582
Female	[0.43]	19.5%
Speech Act	p = 0.315 *
Command	[0.69]	42.3%	26
Declarative	[0.48]	20.6%	34
Question	[0.48]	18.3%	60
Indirect Command	[0.35]	11.8%	17

* These p-values were not selected as significant by the statistical model.

shows the results of this analysis, along with the AIC (Akaike information criterion) score—an indicator of quality of a given model.

From Table 7, it can be observed that speaker gender and interlocutor relationship were significant predictors of 2PS usage. Interlocutor age approached significance. Interlocutor gender and speech act were not selected as significant by the statistical model. The most predictive variable was speaker gender, with female speakers favoring ustedeo while males favored tuteo. For interlocutor relationship, distant interlocutors favored ustedeo and intimate interlocutors favored tuteo. Although only approaching significance, older interlocutors favored ustedeo while same age interlocutors favored tuteo. A summary of native speaker data from Denbaum-Restrepo and Restrepo-Ramos (2022) is presented in

Table 8. Summary of variables that significantly favor the usage of each 2PS for native speakers.

Variable	2PS Favored
Interlocutor Age
Same Age	tuteo
Older	ustedeo
Speaker Gender
Male	voseo, ustedeo
Female	tuteo
Speech Act
Discourse Marker	voseo, tuteo
Question	voseo, tuteo
Commands	voseo
Indirect commands	tuteo, ustedeo
Statement	tuteo
Speaker Age
Young	voseo
Intermediate	voseo, tuteo
Old	tuteo, ustedeo
Interlocutor Relationship
Intimate	voseo
Distant	tuteo (compared to vos), ustedeo
Subject Pronoun Expression
Pronoun and verbal inflection	voseo
Verbal inflection only	tuteo

. In Denbaum-Restrepo and Restrepo-Ramos (2022), three different analyses were carried out (vos vs. tú; vos vs. usted; usted vs. tú). See Denbaum-Restrepo and Restrepo-Ramos (2022) for details of the analyses for the native speaker data.

As can be seen in Table 8, several variables significantly condition the usage of 2PS for native speakers in Medellin including interlocutor age, speaker gender, speech act, speaker age, interlocutor relationship, and subject pronoun expression. For both learners and native speakers, speaker gender and interlocutor relationship significantly constrain 2PS variation. Specifically, for both groups, distant interlocutors favored ustedeo and intimate interlocutors favored tuteo. In terms of speaker gender, learners show the opposite tendency of that of native speakers. For learners, female speakers favored ustedeo, but for native speakers, male speakers favored ustedeo.

4.2. Perception Results

In this section, the perception results from the matched guise task will be presented by adjective (kind, honest, professional, close to interlocutor). An ANOVA was run utilizing a post hoc Tukey test in order to compute pairwise comparisons between learners and native speakers. The mean values for each 2PS for learners will be compared to those of native speakers from Medellin from the Denbaum-Restrepo and Restrepo-Ramos (2022) study. First, results are presented for the personal characteristic of ‘kind’ in Figure 3.

Learners rated ustedeo the highest, slightly higher than tuteo, and they rated voseo the lowest. Native speakers also rated tuteo the highest but ustedeo the lowest, with voseo in the middle. There were no significant differences between learners and native speakers for any of the 2PS (p = 0.530 for tuteo; p = 0.807 for voseo; p = 1.000 for ustedeo).

As can be observed in Figure 4, for the characteristic of ‘honest’, learners rated voseo the highest, followed by tuteo and finally ustedeo. Native speakers rated tuteo the highest, followed by voseo and finally ustedeo. Once again, there were no significant differences between learners and native speakers for any of 2PS (p = 1.000 for tuteo; p = 0.9666 for voseo; p = 1.000 for ustedeo).

As observed in Figure 5, similar to their ratings for ‘kind’, learners rated voseo as the most professional, followed by tuteo and lastly ustedeo. In contrast, native speakers rated ustedeo as the most professional, followed by tuteo, while voseo was rated the least professional. There were no significant differences between learners and native speakers for ustedeo (p = 0.956). However, learners rated tuteo (p = 0.038) and voseo (p < 0.001) as significantly more professional than native speakers. Finally, results for closeness to the interlocutor are presented in Figure 6.

As can be observed in Figure 6, learners rated tuteo as representing the closest relationship with the interlocutor, followed by voseo, which was rated slightly higher than ustedeo. In contrast, native speakers rated voseo as the closest to the interlocutor, followed by tuteo and lastly ustedeo. There were no significant differences between learners and native speakers for tuteo (p = 0.963) or ustedeo (p = 0.605). Native speakers rated voseo significantly higher than learners (p = 0.008).

5. Discussion

The first research question inquired about learners’ frequencies of each 2PS compared to native speakers of Medellin. It was observed that learners’ rates of 2PS production had the same hierarchy as natives from Medellin: tuteo > ustedeo > voseo. However, learners’ percentages differed greatly from native speakers due to a tremendous overgeneralization of tuteo and almost non-existent usage of voseo. These findings corroborate results from previous studies, as Villarreal (2014) and Lyster and Rebuffot (2002) also found an overgeneralization of tú in Spanish and tu in French, respectively. The fact that voseo was practically non-existent shows that it is very slow to be acquired, which can be especially noted given that these learners have been exposed to voseo in Medellin for an average of 7.7 months. Thus, even with this great amount of exposure and input, learners are still barely producing voseo. Ringer-Hilfinger (2012) observed similar results for the acquisition of a phonological variable: the interdental fricative /θ/. She observed only six tokens of /θ/ out of 209 potential contexts—only 2.8%. These six instances of /θ/ were produced by the same two learners. These results are very similar to voseo rates in the present study, as there were only three voseo tokens or 2.1% of all 2PS production by learners. Like in Ringer-Hilfinger (2012), voseo was only produced by two speakers. While this was the case for these two studies, in general, dialectal variants have been reported to be acquired after study abroad programs, especially in longer ones. For example, in George’s (2018) study, learners who studied abroad in central Spain increased their production of vosotros overall over one semester. Lindley (2021) found that many learners who studied abroad in Spain for one year increased their usage of [θ] and [χ]. Likewise, in short-term study abroad programs (7 weeks), Kanwit et al. (2015) and K. L. Geeslin et al. (2010) observed that learners moved toward native speakers in terms of rates and linguistic constraints for the present perfect (Valencia, Spain) and leísmo (León, Spain) respectively. In contrast to these findings, other variables have been found to not be acquired to the same extent. Denbaum-Restrepo (2023b) found that only two out of the six participants increased their rates of wh-question non-inversion after studying abroad in the Dominican Republic for 6 weeks. In addition, Linford et al. (2021) observed that learners studying abroad in the Dominican Republic for one semester did not adopt s-weakening.

The second research question sought to examine what linguistic and extralinguistic constraints condition the use of 2PS for learners compared to native speakers. For learners, the two variables that were found to condition 2PS usage were speaker gender and interlocutor relationship, while interlocutor age approached significance. These three variables were also significant predictors constraining 2PS usage for native speakers. However, in addition to these three variables, speech act, speaker age, and subject pronoun expression were also significant for native speakers. Furthermore, out of the two constraints that were significant for learners, speaker gender showed the opposite trend to native speakers, in that female learners favored usage of usted, whereas male native speakers favored usted. Thus, it appears that these learners’ systems are still developing, as learners have still not acquired all the linguistic and extralinguistic constraints that condition native speaker usage.

It should be noted that they have not yet acquired speech act as a conditioning factor, which is a very important variable for native speaker usage of 2PS. As their proficiency and exposure increases, learners should start to acquire more constraints to pattern more like the native speakers. Previous research on other variables has found that as time spent studying Spanish increases, the linguistic constraints that condition the use of variable structures start to pattern more like those of native speakers. For example, K. Geeslin et al. (2015) found that learners’ linguistic constraints conditioning subject expression did not match those of native Spanish speakers until the graduate student level, with fourth-year students starting to approach native-like patterns. In addition, Kanwit (2017) observed the same pattern of linguistic constraints lining up with respect to those of native speakers for the L2 acquisition of future-time reference. Kanwit (2018) describes an underlying rule system that is dynamic and changing over time. Previous research has found that pragmatic constraints are acquired later than linguistic ones (K. Geeslin, 2003; Woolsey, 2008). For example, Woolsey (2008) found that pragmatic factors, such as frame of reference (comparing someone to themself) and experience with the referent, were acquired late. In fact, frame of reference was acquired so late that even learners of the highest proficiency level had not acquired it (i.e., use estar in these situations). Given these findings from previous research, it makes sense that learners take longer to acquire speech act, since it is a pragmatic constraint. Learners’ systems are dynamic and changing, including non-native-like generalizations of certain structures. In the earlier stages of development, a learner’s system reflects variation that shows different patterns than those of native speakers. We see this both with rates of 2PS with the learners in this study as well as with their developing systems, demonstrated by the linguistic and extralinguistic constraints. Learners do, however, pattern like native speakers for interlocutor relationship and interlocutor age, in that distant and older interlocutors favored ustedeo. In terms of the variable speaker gender, for learners, it was found that females favored ustedeo, whereas for natives, it was observed that males favored ustedeo. This could be due to that fact that for the learner group, the average proficiency level of their current Spanish course was slightly higher for females than for males (females = 8.67; males = 7.67). Previous research has not necessarily observed this before, but it is an expected association.

The third research question sought to examine learners’ perceptions of each 2PS compared to that of native speakers. Findings showed that learners’ perceptions toward each 2PS differ from native speakers, but only to a significant extent for two attributes. For the characteristic of ‘kind’, learners rated ustedeo the highest, and for ‘honest’, they rated voseo the highest. In contrast, native speakers rated tuteo the highest for both adjectives. However, it should be noted that there were no significant differences found between learners and native speakers for ‘kind’ and ‘honest’. For ‘professional’, learners rated voseo highest, whereas native speakers classified ustedeo as the most ‘professional.’ Last, for ‘closeness’, learners had the following hierarchy: tuteo > voseo > ustedeo. Meanwhile native speakers’ evaluations were as follows: voseo > tuteo > ustedeo. Although learners show some differences from native speakers, in the majority of cases, there were no significant differences. For closeness to the interlocutor, learners are still not aware that voseo is perceived as the closest 2PS by native speakers. This finding has a similar connection to many sociolinguistic variables previously studied in L2 French and Spanish, as learners are required to establish close relationships with locals, as measured by the social networks created, in order to develop native speakers’ rates of production (Kennedy Terry, 2017, 2022; Trentman, 2017). Moreover, the role of proficiency goes in line with close relationships and social networks, as found by Pozzi (2021). However, learners do perceive voseo as more intimate than ustedeo. Thus, although learners have not fully acquired local nuances of 2PS—they have partially acquired them. These findings resonate with previous research (Grammon, 2024) that has dealt with L2 evaluation of the appropriateness of dialectal variants linked to a specific speech community.

Putting these results together with findings from usage data from the DCT, we can conclude that although learners perceive vos to a somewhat native-like extent and the role that it plays in the local variety, learners do not actually employ voseo. This lack of usage of voseo could be due to various reasons. First, vos is a very local variant—a strong identity marker for Paisas (Jang, 2013). Even though learners are immersed in the culture and the language for extended amounts of time, they may need even more time to acquire it. Specifically, they may need more time to feel a part of the Paisa culture and to acquire a sort of Paisa identity. Second, another possible explanation is that it is possible that even though learners are aware of voseo, they might not want to acquire it. Even though they are living in Medellin, they might choose to not identify with Paisa Spanish for various reasons. Pozzi (2021) affirms that the amount of time a learner spends with their local social network corresponds to a strong predictor of voseo use at the end of a study abroad program. Along with other factors such as the type of task and mood, the effect is higher in high-proficiency learners. Moreover, it has been shown that learners do acquire sociolinguistic variation during their studies abroad, including in countries where languages other than Spanish are spoken officially. Kennedy Terry (2017, 2022) also found that some French phonological variation patterns are significantly predicted by a higher score in their social network scale, which shows the predictive power of social networks in the acquisition of sociolinguistic variables in French. The effect of social networks was also suggested in English-speaking learners of Arabic during a study abroad program (Trentman, 2017). In her examination of the interdental fricative with Spanish learners studying in Spain, Ringer-Hilfinger (2012, p. 442) affirms that “learners make a conscious choice to use the variety that best suits their needs”. Similar to the present study, learners may not feel that they need to use vos in order to survive in Medellin, and they might consciously decide not to use it. Lastly, vos is a new form for learners and has its own verbal conjugations. It might seem too complex for learners at this point in their acquisition process. Furthermore, voseo might not be included in their explicit language instruction, as Spanish language classes often do not cover instruction of voseo, at least in US classrooms (Shenk, 2014).

6. Conclusions and Pedagogical Implications

The present study lays a foundation for examining the L2 acquisition of a tripartite 2PS system in the study abroad context. It further elucidates previous research on the L2 acquisition of 2PS, since tripartite systems are more complex and have many more nuances to acquire, as is especially the case of 2PS in Medellin due to the dual usted (Denbaum-Restrepo & Restrepo-Ramos, 2024). The current study incorporates language perceptions into the study of the L2 acquisition of 2PS, which allows for further insight into the usage of 2PS. Results showed that learners produce 2PS in the same hierarchy as native speakers but in a much different way, since they overgeneralize tuteo and barely produce voseo. In addition, learners’ 2PS usage is constrained by two variables: speaker gender and interlocutor relationship. Finally, learners’ perceptions of 2PS are similar to native speakers for the attributes of ‘honest’ and ‘kind’ but show some significant differences for ‘closeness to the interlocutor’ and ‘professional’. Altogether, we can conclude that although learners perceive vos and the role that it plays in the local variety to a somewhat native-like extent, learners do not actually use it.

Findings from the present study have pedagogical implications. As language instructors, we should expose learners to sociolinguistic diversity starting in beginning language classes. For example, as part of our language courses, we can expose learners to various Spanish dialects and various contexts of use for each one, including informal and formal contexts. There are several ways this can be achieved, including inviting guest speakers to come talk with students in the class, watching videos or television series, or organizing virtual conversation partners with students from different countries. In addition, sociolinguistic data can be presented to students. This would include both teaching basic dialectology to students, which would aid in comprehension, and explaining to students the idea of linguistic stereotypes and the fact that the way one group of people speaks is not actually indicative of their intelligence level. Finally, practicing listening comprehension of various dialects could result in an increase in positive attitudes toward those dialects (see e.g., K. L. Geeslin & Schmidt, 2018).

While the present study examined the L2 acquisition of 2PS in Medellin at one point in time, it would be illuminating for future studies to collect data longitudinally in order to observe the development of learners’ 2PS usage and perception over time and how it changes from before they have exposure to monodialectal input throughout their stay. While the present study lays the foundation for the study of the L2 acquisition of 2PS in the study abroad context, our participants consisted of a very heterogeneous group, including countries of origin, L1s and L2s, Spanish level, and time living in Medellin. It should be acknowledged that the current study is quite limited due to the number of participants and the fact that factors such as proficiency and time/exposure are only indirectly accounted for. Therefore, future studies should attempt to investigate the L2 acquisition of 2PS with a more homogenous group of learners and with a larger number of participants. The present study examined learners’ perceptions of 2PS, but it would also be productive for further studies to examine other individual differences, such as social interaction with native speakers and learners’ identities and attitudes toward Paisa culture. It could also be interesting for future studies to examine learners’ understanding of voseo and the extent to which it is included in language instruction, whether that be in learners’ countries of origin or in Medellin. It is our hope that the present study serves as a reference for future work on the acquisition of vos as a sociolinguistic variable, while fostering additional studies on this phenomenon in Paisa Spanish and elsewhere.