Structural Variation in Northern Peru: The Case of the Estar + Que Periphrasis

1. Introduction

In northern Peru, a regional variant is observed in Spanish that describes actions in progress. This structure is used in many of the same semantic contexts as the present progressive or imperfect past progressive and can be schematized as estar (V₁) + que + verb (V₂), in which V₂ is conjugated in the same person, number and tense as estar (V₁). Examples of the construction from observed speech are given in (1):

(1) a. Tu papá está que te llama.
b. Estaba que te miraba.
c. Están que se bañan.

This verbal string is used in place of the normative estar + present participle (e.g., estoy comiendo, I am eating) and previous research limits it strictly to concomitant actions. Spanish has no resource for distinguishing between concomitant events and durative (or habitual) events (I am eating now versus I am eating a lot lately) and thus present progressive (PROG) is utilized for both aspectual contexts (Arrizabalaga, 2010, p. 15). According to Arrizabalaga, the estar que periphrasis^[1] fills this void, thus disambiguating the aspectual lens of the utterance. The examples in (2), adapted from Arrizabalaga (2010, p. 16), compare the semantic use of the estar que periphrasis and the progressive (emphasis mine):

(2) a. Está estudiando para el examen del lunes.
beprs.3sg study presp for the test of the Monday
‘(He/she) is studying for Monday’s test.’

b. Está que estudia para el examen del lunes.
beprs.3sg relthat studyprs.3sg for the test of the Monday
‘(He/she) is studying for Monday’s test.’

c. Está estudiando tercer año de Derecho.
beprs.3sg studypresp third year of Law
‘(He/she) is studying [his/her] third year of Law.’

d. *Está que estudia tercer año de Derecho.
beprs.3sg relthat studyprs.3sg third year of Law

Examples (2a) and (2b) describe an action that occurs concomitantly with the speech act and are both considered grammatical in the northern Peruvian variety. According to current literature, of the latter two examples that describe ongoing or durative actions, only (2c), the progressive, is grammatical. Unlike the progressive form, which has extended to include some habitual aspects (Arrizabalaga, 2010; Márquez Martínez, 2009; Merma Molina, 2008; Torres Cacoullos, 1999, 2012), the estar que construction is understood as being restricted to describing only actions occurring simultaneously with the spoken utterance or with the referenced point in the past (Arrizabalaga, 2010, p. 71). The construction in question can be seen in both present and past imperfect tenses as well as in all person conjugations, as seen in examples (3) and (4). Example (3a) demonstrates use in the present tense of third person singular (3SG) and (3b) demonstrates the construction in the imperfect past 3SG, both used to describe actions concomitant to the speech act or reference point.

(3) a. Está que come.
beprs.3sg relthat eatprs.3sg
‘She/he is eating [right now].’ /* ‘She/he eats [in general].’

b. Estaba que comía.
bepst.ipfv.3sg relthat eatpst.ipfv.3sg
‘She/he was eating [at that time].’ / *‘She/he was eating [lately/every day, etc].’

To provide more examples, (4) presents instances of the estar que periphrasis observed by the author in spoken conversation in Trujillo, Peru in June of 2017.

(4) a. Está que llueve.
beprs.expl.3sg relthat rainprs.expl.3sg
‘It is raining.’

b. Recién estoy que almuerzo.
recent beprs.1sg relthat eat lunchprs.1sg
‘I am just now eating lunch.’

c. Están que lo preparan.
beprs.3pl relthat it prepareprs.3pl
‘They are preparing it.’

d. Están que se bañan.
beprs.3pl relthat refl.3pl batheprs.3pl
‘They are swimming.’

All of the examples provided in (4) were spoken to describe events in action at the time of speech. In these examples, we observe that the two verbs (V₁ and V₂) are conjugated in the same tense and person. Thus, a statement like *Está (3SG) que almuerzo (1SG) is ungrammatical and not observed in speech in the northern Peruvian variety.

The current study aims to describe the variable use of this construction and analyze its structural and aspectual characteristics based on analysis of analogous structures and new usage data. The analysis is viewed through the lens of language variation and grammaticalization, building on work done by Arrizabalaga (2010), who has to date provided the only extensive analysis of what he terms the perífrasis concomitante norperuana ‘northern Peruvian concomitant periphrasis.’ Concomitant, for the purposes of this paper, is taken to mean that the occurrence of the described event is simultaneous with the speech act or, in the case of imperfect past tense, with the point of reference (Cohen, 1993). The specific goals of this paper are to add to the knowledge of the estar que construction by 1) demonstrating a broader and more variable use than previous research was able to describe and 2) analyzing the construction’s clausal structure as compared to analogous structures. Complement placement is tested as a potential point of variation for analysis of the clausal structure.

2. estar + que and verbal periphrases in Spanish

Arrizabalaga (2010) introduces the estar que construction as a reaction to grammaticalization of the gerund progressive (PROG). PROG has extended to include aspects of habituality and durativity along with concomitance; thus, the estar que construction developed as an unambiguous concomitant structure. This section lays out the aspectual and structural restrictions that previous literature has placed on the phenomenon. It also introduces structures of potentially analogous origin for the estar que construction.

2.1. Aspectual and Structural Restrictions on the Northern Peruvian Concomitant Periphrasis

Arrizabalaga (2010) places both aspectual and structural restrictions on the estar que structure. Aspectually, it is only compatible with concomitant, progressive aspect and is not observed nor compatible with habitual or durative aspects. Structurally, he observes that it is uncommon to find the phenomenon with negation and that if it does occur, it only does so before the second verb in the sequence (V₂). This restriction is connected to the structure’s concomitant aspect: if the construction is limited to actions that are happening simultaneously with the speech act, it follows that it is counterintuitive to negate these actions, as that would require the action to not be happening. The progressive in Spanish also passed through a phase of not allowing negation (Arrizabalaga, 2010, pp. 102, 194), which helps support the idea of the estar que construction following a similar grammaticalization path as the PROG construction. In terms of structure, Arrizabalaga assumed a monoclausal syntax for the estar que periphrasis, following the general PROG construction of auxiliary estar + lexical gerund. Additionally, like negation, clitic pronouns are described as only appearing before V₂. The examples in (5) demonstrate these restrictions (Arrizabalaga, 2010, p. 17). New data expands previous research’s strict limitations on structure and aspectual use, showing that speakers accept variable aspectual use and complement (negation and clitic pronoun) placement.

(5) a. está que te busca/estaba que te buscaba
b. *no está que te busca
c. *te está que busca

The examples in (5) demonstrate what we previously knew about the acceptable placement of clitic pronouns (in this case, te), found preceding V₂ (5a), as opposed to the ungrammatical pre-V₁ placement (5c). (5b) demonstrates the ungrammaticality of negation with this phenomenon. For clarification of estar que usage compared to its PROG counterpart, Table 1 summarizes both constructions’ aspectual and structural restrictions (Arrizabalaga, 2010). Examples of each category are included.

To summarize, the estar que construction developed as a more precise option for discussing concomitance in the face of PROG’s aspectual extension. Thus, it is more restricted than the traditional PROG, not accepting simple past, stative verbs or habitual aspect and with limited use of clitics and negation. However, participants in the current study accepted clitic placement and negation outside of these previously proposed restrictions, evidencing variation and fluctuating use of the construction. Before reporting these data, I present analogous structures that could provide insight to the clausal structure of the estar que construction.

2.2. Comparison Structures for Analogy

In analyzing the clausal structure of the estar que construction, I compare two constructions as potential sources for origin through analogy: attributives and iteratives, seen in (6) and (7), respectively. In both (6) and (7), estar is used with another verb or set of verbs that are conjugated for the same person, number and tense as estar, much like the estar que periphrasis in question.

(6) está que trina
lit. (he/she) is that (he/she) trills
‘(He/she) is angry.’

(7) está habla y/que habla
lit. (he/she) is (he/she) talks and/that talks
‘(He/she) is talking/has been talking a lot.’

Structures like those in (6) are commonly found in many dialects of Spanish (Arrizabalaga, 2010; Lisyová, 2004; Penadés Martínez, 1991), but are not productive in northern Peru, as observed by the author and corroborated by Arrizabalaga (2010). The attributive structure in (6) functions as a reduced form of an attributive statement with an adjectival complement: Ella está [tan enojada] que trina, with estar functioning as a lexical verb describing a state that can take complements on its own: está que trina = está molesta = lo está (Arrizabalaga, 2010, p. 190). This statement qualifies the subject ella and can be replaced with an adjective without losing its semantic meaning: Ella está muy enojada ‘She is very angry.’ While the attributive está que trina contains a level of concomitant aspect, it is only found with a figurative reading (i.e., ella is not literally trilling). This is an expression used to attribute a state of being rather than to describe a simultaneous action. Semantically, this construction and the estar que periphrasis achieve different goals: the attributive expresses a mood or attribute while the estar que periphrasis expresses an occurring action. However, the structures appear very similar.

Iteratives (7) also appear similar in surface structure to the estar que periphrasis and are, indeed, prevalent in the northern Peruvian region under study. Treviño (2004) discusses iterative structures in Spanish, which include a repetition of the main verb, lending an aspect of a repeated or repetitive action. Treviño comes to the conclusion that the compound verb_i + y + verb_i, (as in 8) is a fixed expression that only functions with the auxiliary estar, much like the present participle (gerund) or infinitive: está comiendo, ‘he/she is eating’ or está para comer, ‘he/she is about to eat’. Example (8), taken from Treviño (2004), provides observations of this structure with evidence of its fixed nature noticeable in the use of the unmarked singular and often conjugation with vowel –e, regardless of thematic vowel, tense, or person and number, as seen in (8b, 8c, 8c) (bolded emphasis mine):

(8) a. Está lee y lee ese poema de Blake (154).
b. Ha estado pregunta y pregunta por usted (151).
c. Ahí estaban toque y toque en honor…de “México Posible” (152).
d. …arme que arme escándalos (158).

In (8), while the two repeated verbs match in person and number, they do not necessarily conjugate for the subject of the sentence. Considering this structure, it follows that while attributives like (6) are biclausal — composed of two lexical verbs —, iteratives like (7) and (8) are monoclausal — composed of an auxiliary verb and a fixed string of main verbs. These structures are schematized in (9a) and (9b) and compared to the structure of PROG with a gerund (9c).

(9) a. [_IP[_VPEstá_i] [_CPque [_VPtrina_i]]] attributive
b. [_IP[_AuxPEstá [_VPlee y lee [_NPese [poema]]]]] iterative
c. [_IP[_AuxPEstá [_VPleyendo [_NPese [poema]]]]] PROG-gerund

The monoclausal structures in (9b) and (9c) are more analogous with the progressive and Arrizabalaga’s classification of estar as an auxiliary verb; however, (9a) provides motivation for the person agreement between the two verbs that is seen in the concomitant periphrasis, as opposed to the fixed construction of the monoclausal iterative (7, 8) or PROG. Due to the uncommon agreement of person and the variable acceptance of proclisis before both V₁ and V₂ found in the current data (described in subsequent sections), this study leans toward an understanding of the estar que construction as biclausal, deviating from Arrizabalaga’s claim on the auxiliary status of estar. While semantically comparable to the PROG construction, the estar que periphrasis functions under unique restrictions. In order to explore use and variation, restrictions of habituality and complement (clitic pronoun and negation) placement are the principal focus of the present study. Evidence of broader use and variation of this construction allow us to speculate that it is a construction in flux.

3. Methodology

3.1. Research questions

In order to expand on previous research and clarify geographical distribution and usage restrictions, an acceptability study was carried out in Trujillo, Peru and Machala, Ecuador during the months of June and July in the year 2017. As mentioned above, Arrizabalaga’s (2010) work on the northern Peruvian concomitant construction is the only known in-depth study of the estar que construction. His study is based on data from Piura, Peru, a coastal region in the extreme northwest of the country. He determines that the phrasal structure is attested only in the very northernmost coastal region of the country. In addition to discussing the aspectual use and structural makeup of this phenomenon, I expand on the understood geographical extent of this phenomenon. This study aims to answer the following research questions:

1. Geographically, to where does the use and acceptance of the estar que construction extend?
2. In what aspectual uses is the estar que construction accepted? What does this use say about the stability and/or grammaticalization of the phrase?
3. What are placement restrictions on the estar que construction in terms of negation and clitic pronouns and what do these restrictions evidence about the clausal structure?

The estar que construction is an extremely colloquial structure that carries a measure of stigma. Arrizabalaga notes that there is “strong resistance to [the construction’s] use in formal registers” (2010, p. 17) and his participants from Piura commented that while the structure is very frequent, “es incorrecto” and “no es una conjugación correcta” (2010, p. 76). As Arrizabalaga (2010) notes, these types of colloquial phenomena are difficult to find documented in written sources. Examples of this construction were not found in online corpora^[2]. Thus, in order to confirm the claims made by Arrizabalaga and investigate the geographical distribution of the estar que periphrasis, acceptability questionnaires and conversation observations were carried out in Trujillo, Peru and Machala, Ecuador. Trujillo is located 262 miles south of Piura, the city in which previous research was conducted; Machala is 47 miles north of the Peruvian border. The choice of Trujillo as a research site was intended to explore a southern reach of the phenomenon. Although Arrizabalaga explicitly states that se extraña, ‘it is strange’ to hear the estar que construction as far south as Trujillo (Arrizabalaga, 2010, p. 80), the author had observed it there on multiple extended visits to the city in 2010, 2011 and 2013, sparking the interest in the present study. Machala, Ecuador was chosen to document use or recognition of the construction north of previously documented use. Even as this phenomenon is known only as a Peruvian expression, political boundaries do not necessarily line up with dialectal boundaries; there is little reason to believe that language documented 50 miles south of the border with Ecuador would not be observed in southern Ecuador as well. However, the only reference to usage across the Ecuadorian border in previous research is that “Sabemos que resulta extraño al otro lado de la frontera, en las provincias del sur del Ecuador” (Arrizabalaga, 2010, p. 80). In order to approximate the degree of use of the construction, an acceptability questionnaire of 60 items was given to 35 participants in Trujillo and 14 in Machala.

3.2. Participants

Surveys were administered in private residences or places of work. The birthplace of the participant and that of his/her parents were included in the demographic portion of the questionnaire in order to ensure that participants’ speech was developed in and influenced by the region under study as opposed to varieties from Piura. While an attempt was made to include participants of various socioeconomic statuses and varying age ranges, no specific quota was followed for any group.

Of the 35 participants from the Trujillo data set, 27 had a university-level education or were in the process of obtaining one at the time of the survey. Three of the participants had completed a degree beyond a university level and three had completed a high school level education without further studies. Two had an elementary level education. Twenty-one participants were female and 14 were male. The average age of the participants was 34.5, ranging from 18 to 68.

The Machala data set included 14 participants. Of the 14, 12 responded regarding level of education. Of these 12, eight had a university-level education, three had a high school education and one had completed an elementary level education. Ten of the Machaleños were female and four male. The average age of the Machala participants was 41.4, ranging from 18 to 63. Less participants were found for this portion of the study, as neither the author nor previous literature had encountered the structure in this region. The goal of this set of data collection is mainly to establish use and/or acceptance, which could be done with fewer participants than in Trujillo, where the goal was to understand nuances of a structure already observed.

3.3. Questionnaire and Interview

In a stimulus-driven elicitation exercise, participants were asked to describe the actions portrayed in images presented to them on paper (TFS Working Group, 2011), with the goal of eliciting spontaneous use of the concomitant periphrasis. However, only four examples of estar que — three from the same participant — were organically produced during this task, so this portion will not be analyzed in the current paper. That few cases were provided spontaneously during the interview is in line with its status as a non-standard construction that may not present in a semi-formal interview context^[3].

The acceptability questionnaire is comprised of 60 sentences, 40 of which contain the construction under study, 20 of which are distractors (examples are presented in Table 2). The sentences were created by the author based on Arrizabalaga’s elicited written responses to questions regarding the structure in “informal surveys” (Arrizabalaga, 2010, p. 75) and on the author’s own observed examples from time spent in the region and continued communication with inhabitants of Trujillo. Sentences were created in two categories: those hypothesized to be acceptable and those hypothesized to be unacceptable, based on the restrictions outlined by Arrizabalaga in Section 2 and Table 1.

This construction is a well-known regional variant that prompted remarks evident of social conditioning in follow-up conversation after completing the questionnaire. Many participants noted that people from outside the city center — the notably less economically stable areas — might use the construction often, but they themselves do not. Even participants whom the author observed using the estar que construction in casual speech commented that they don’t use it. I was often instructed to go to an area of the city with residents of lower socioeconomic status in order to “find” my variable in the wild. Eliciting stigmatized variants often results in data that is skewed toward standard speech. This was evident in the lack of examples provided organically in the picture-describing task. If the estar que construction is socially marked, we can assume any use (or acceptance) present in the data points to more use in spontaneous, informal speech.

The items included in the questionnaire were paired with a Likert-style scale from one to five:

(1) – No es aceptable y no lo dice nadie.

(2) – No es aceptable, pero sí se dice.

(3) – No estoy seguro/a.

(4) – Es aceptable, pero yo no lo diría.

(5) – Es aceptable y yo lo diría.

Answers of 2, 4, and 5 were all counted as acceptable. The present scale is unique in that (2) is included as “acceptable,” as it indicates recognition of use of the construction, although not necessarily the judgement of being grammatically “correct.” Giving the participants qualifying attributes in addition to number rankings allowed participants to mark what would most closely resemble actual usage and not solely social opinions or judgments. Thus, a participant could mark that they had heard the construction used by others, even though he/she does not agree that it should be used or is grammatically correct. Option (3) allowed for participants to not feel pressure to definitively know a rating by answering ‘I’m not sure.’

Based on previous research and the author’s own observations, it was hypothesized that the estar que periphrasis would be accepted with any dynamic verb in which the present progressive could be used with a concomitant aspect. Following previous research, it was hypothesized that only the V₂ would select complements; thus, in sentences with a direct object pronoun or reflexive pronoun, it was hypothesized that these complements would only be accepted preceding V₂. I hypothesized a low acceptance of negation in general, with slightly more acceptance pre- V₂ than pre- V₁. It was also hypothesized, in accordance with previous research’s restrictions of aspect, that the construction would not be accepted with habitual meaning (Arrizabalaga, 2010). This study also includes examples gathered through informal conversation and observation.

4. Data and Analysis

Figures 1 and 2 demonstrate the overall percentage of acceptance among the Trujillano participants and the Machaleños, respectively, for the various elements studied and compare them to the base structure, which is taken to be the estar que + verb construction in the present tense without negation or pronouns and referring to a concomitant action. This is considered the base structure because of its hypothesized acceptance based on Arrizabalaga’s study and the author’s previous anecdotal observations. An example of a base structure sentence is Mi mamá está que cocina.

Figure 1.Percentage of acceptance of estar que construction by category: Trujillo

Figure 2.Percentage of acceptance of estar que structure by category: Machala

“Acceptance,” as already noted, includes responses of 5, 4 and 2; however, in the figures above, I separate this acceptability by category of responses 2 and another of 4 and 5 in order to better visualize levels of acceptance versus use. Responses of “2” indicate the participant admits use by others while making a negative judgment about that use; responses of “4” and “5” point toward grammatical acceptance regardless of individual personal use. The data in Figures 1 and 2 demonstrate the level of acceptance by overall percentage of answers of the Likert-style scale.

The base structure is accepted at 87% in Trujillo and 81% in Machala, evidencing a high general familiarity with the construction in both regions. Object clitic placement pre-V₂ (hypothesized acceptable) was accepted at 87% in both Trujillo and Machala. These charts show a clearly higher acceptance of this placement of clitic pronouns as opposed to pre-V₁ (51% in Trujillo and 74% in Machala), which was anticipated.

Negation shows unexpectedly high acceptance rates in both regions, with an unanticipated pattern in Trujillo and the anticipated pattern in Machala. Trujillanos accepted more pre-V₁ placement (at 70.5%) than pre-V₂ (at 61%), which was unanticipated. Machaleños accepted only 58.5% of pre- V₁ negation but accepted more than 85% of pre- V₂ negation. Habitual readings were also accepted more than hypothesized. Machaleños accepted more habitual use than did Trujillanos: 81% acceptability in Machala, 72% in Trujillo, both of which are higher than anticipated, considering previous assumptions of concomitance restrictions.

5. Discussion

For the purposes of the aspectual and structural analysis of this paper, the discussion is focused on use of habitual aspect and negation and clitic pronoun placement. However, I note that, as expected, verbs in the simple past are not accepted in the estar que construction and that stative or emotive verbs — hypothesized to be incompatible — show some acceptance at close to 50% in Trujillo and 68% in Machala. Rejection of the simple past upholds the assumption of progressivity, while some acceptance of stative/emotive verbs demonstrates a level of syntactic variation.

5.1. Habituality

Examples with habitual readings were accepted at a higher rate than hypothesized (72% accepted in Trujillo and 81% accepted in Machala), as seen in Figures 1 and 2. Examples observed in casual speech in Machala (10a) and Trujillo (10b) that convey habituality are provided in (10).

(10) a. O sea, contigo está que te da todo el director.
‘That is, the director is giving you everything [lately].’

b. …y siempre están que se queman la lengua.
‘…and they’re always burning their tongue [on the tea].’

While two examples are not strong evidence, finding observations of habitual use in spontaneous, casual speech provide a lead to evidence of a more general habitual use of the construction. Due to the generally high acceptance of habitual uses on the acceptability questionnaire, the construction is viewed as potentially representing a move toward inclusion of habitual use. This extension of aspect is much like that which the PROG construction has undergone, seen in the grammaticalization cline in (11) (Torres Cacoullos, 1999, 2012). The PROG construction has developed from originally purely locative to progressive to more recently including habitual or durative aspects; it is possible the estar que construction is following a similar pattern or contains a gradient reading that easily passes into habituality.

(11) ESTAR + -NDO [PROG]

locative construction >
a. Está en Managua, en Nicaragua, visitando

progressive aspect >
b. Uno nunca puede ver en la cara del juez lo que está pensando
c. estoy pensando en mi hija

habitual/durative aspect
d. Lo que están haciendo es invertir en infraestructura cultural
f. Está haciendo el doctorado en una universidad privada

The acceptance of habitual use of the estar que construction provides evidence that this construction is not exclusively limited to describing concomitant or progressive actions. Highly accepted examples from the current questionnaire data are used in (12) to form a cline parallel to Torres Cacoullos’ (1991) above.

(12) ESTAR QUE + VERBO

progressive concomitant aspect >
a. Mi mamá no puede hablar ahorita; está que cocina.
b. No puedo salir en este momento; estoy que estudio.

habitual/durative aspect
c. ¿Estás que trabajas todos los lunes?
d. Antonio está que visita a su mamá todos los fines de semana.

The examples in (12c) and (12d) include phrasing (todos los lunes/los fines de semana, every Monday/weekend) that clearly marks habitual activities while (12a) and (12b) clearly refer to concomitance with the phrases ahorita and en este momento. In comparing the cline (11) and (12), I suspect that the estar que phenomenon may be following a similar grammaticalization pattern to that the PROG construction has already undergone.

5.2. Negation, Pronouns and Clausal Structure

Figures 1 and 2 in Section 4 include the acceptance of negation and pronoun placement in the estar que construction, yielding unanticipated results according to the original hypotheses of the questionnaire. Clitic pronouns loosely follow the hypothesis of being preferred preceding V₂: in Trujillo, there was 87% acceptance pre-V₂, versus only 51% accepted before V_1, (where acceptance was not anticipated). In Machala, the rate for the clitic placement before V₂ (84%) was comparable to that of Trujillo; however, the acceptance for pre-V₁ placement was much higher, at 74%. This could point to one of two patterns: 1) a wider fluctuation of structural makeup in Machala or 2) less familiarity with the construction in this area. If participants are not intimately familiar with the construction, they may impose less syntactic or semantic restrictions on its use. Although many Machaleños conceded that they had heard the construction in their area and understand it to be used in place of the progressive, most described it as a Peruvian saying. If the participants do not frequently encounter the construction, but loosely understand it as a stand-in for the progressive, accepting clitics pre-V₁ may be an analogy to the present progressive. Overall, higher-than-anticipated pre-V₁ acceptance shows a tolerance for varied pronoun placement.

In regard to negation, the results show a much more ambiguous acceptance than previously hypothesized, ranging from 58-85% average acceptance. The example from the questionnaire that most consistently received high rates of acceptance is the following: La estoy llamando, pero está que no me contesta. This received 93% acceptance in Machala, and only 46% in Trujillo. As a construction connoting concomitant aspect, it is understood that the event or action is taking place at the moment of speech in the present tense or at the point referenced in the past tense. Thus, it would be difficult to discuss the absence, or negation, of an action concurrently taking place. However, the above example, …está que no me contesta, can be viewed as a deliberate action of not answering [the phone], or rather, that not answering is an actionable decision itself rather than a negation of an action. In general, negation seems to be a less straightforward element to use with this construction, explaining the generally lower acceptance in both positions, excepting the high acceptance of pre-V₂ negation in Machala.

The higher-than-anticipated acceptance for clitics and negation preceding V₁ (clitics: 52% average in Trujillo and 74% in Machala; negation: 70% in Trujillo and 59% in Machala) provides conflicting evidence for the clausal structure of the construction. Arrizabalaga understood estar to be an auxiliary verb; however, other monoclausal constructions, such as PROG and iteratives (Treviño, 2004), only allow for negation and clitic placement specifically before the auxiliary V₁ (i.e., No está comiendo, *Está no comiendo; Está comiéndolo, *Está lo comiendo). These constructions are bound phrases that take complements as an entire phrase (i.e., at the beginning or end of the phrase, but not in between verbs), while the estar que construction accepts and seems to favor complements preceding V₂, which seems to indicate a different structure. The variable results of negation and clitic placement within the estar que construction point toward a construction in flux, variably accepting complements in pre-V₁ and pre-V₂ positions. In the following paragraphs, I compare the estar que phrasing to the previously described attributive and iterative constructions, with a focus on the clausal structures.

In Section 2.2, iteratives and attributives were introduced as a means of comparing clausal structure to that of the estar que periphrasis. Following Treviño’s (2004) analysis, the iterative is a monoclausal structure. The syntactic structure of iterative constructions is reproduced below in (13).

(13) [_IP[_AuxP Está [_VP lee y lee [_NPese [poema]]]]]

It has also been observed that attributive constructions (6) are comprised of two clauses: estar as a lexical stative verb and the V₂ as a lexical verb used to describe a metaphorical attribute of that state. The structure of the biclausal attributive is reproduced in (14). While Arrizabalaga (2010) claims that estar in the estar que construction does not function on its own lexically, it appears to do so in attributive statements, which are very similar to the estar que construction.

(14) [_IP[_VP Está_i] [_CP que [_VP trina_i]]]

The Peruvian estar que construction consists of two finite verbs connected by the conjunction que, like the attributive seen in (14). No other periphrasis in Spanish is composed of two finite verbs. For example, the progressive, to which estar que is compared because of its concomitant aspect, is composed of a finite auxiliary, estar^[4], and a nonfinite verb (the present participle (gerund)). Additionally, if Arrizabalaga’s structure of auxiliary + finite verb were true, his conclusions of clitic pronoun placement preceding V₂ would be contradictory to other auxiliary forms in modern Spanish, where clitic climbing to the auxiliary is obligatory (se está bañanado, but *está se bañando, compared to *se está que baña, but está que se baña).

The iterative structure presented in (7) and (13) by Treviño (2004) is highly productive in the northern coastal Peruvian region studied, while the attributive seen in (6) and (14) is not. Thus, it is sensible to surmise that the prevalence of the iterative construction could have made the estar que string easily accepted due to its similar surface structure of two conjugated verbs. However, the estar que construction seems to have developed differently than the iterative: as a biclausal element. Since frequency plays a large role in grammatical change (Bybee, 2001, 2003), the frequency of constructions with estar and of the conjunction que in general in Spanish, as well as the frequency of the iterative construction would have made analogical grammatical change more probable and accessible. If the iterative is a direct originator of the estar que periphrasis, the iterative reading may have bleached out, along with one of the repetitions of the verb; although Treviño’s reading of the iterative construction as a fixed form, does not make the dropping of one of the verbs particularly plausible.

While the attributive construction is not, in fact, productive in this area of Peru, the concomitant periphrasis seems to mirror more closely the syntactic structure of the attributive está [tan enojada] que trina. The concomitant may likewise be a reading of state: está [en el estado] que estudia. Thus, the potential structure for the estar que periphrasis may be the same as that of attributives (15):

(15) [_IP[_VPEstá_i [_CPque [_VPcome_i]]]

This syntactic structure is opposed to Arrizabalaga’s (2010) argument that estar functions as an auxiliary in the estar que construction. Due to the person agreement between V₁ and V₂, I have chosen to represent the construction not as a true periphrasis, but rather in a manner parallel to the attributive structure. A biclausal structure allows for the variable placement of complements, including the more common placement before V₂ compared to V₁. However, if complements of V₂ are able to “climb” to the higher of the two verbs — as is seen in the acceptance of placement before V₁ — we could view the structure as functioning as monoclausal, such as general PROG. Thus, while I represent the estar que construction as essentially biclausal, the ambiguous acceptance of complement placement points to an instable structure. My representation is not meant to provide a definitive syntactic structure, but rather begin initial conversations about the syntax and use of estar que.

6. Conclusions and Further Research

This study explored and analyzed the use and structure of the estar que construction prevalent in northern Peru. The structure was found to be accepted by participants outside of the geographical area defined by previous research, both further south and across the northern border with Ecuador. Based on previous findings and the author’s own observations of the construction in the area of Trujillo, it was hypothesized that use of the concomitant periphrasis would be accepted with any active verb in which the present progressive could be used with a concomitant aspect, while negation and clitic complements were hypothesized to have more restricted use. However, examples of habitual aspect as well as variable clitic and negation placement were accepted at higher-than-hypothesized rates. This variable acceptance evidences a structure in flux, perhaps following a similar grammaticalization path as followed by the present progressive in Spanish. Structurally, I have shown the estar que construction to most likely be biclausal, similar to attributives, but acknowledge that iteratives, a more common construction in northern Peru, may have been an initial analogous structure for estar que’s development.

The written acceptability questionnaire has limitations for rating a predominantly informal oral construction in a formal format. In further research, other methods of data collection would be beneficial for analysis. As an example, audio recordings of the estar que periphrasis could accompany the same acceptability judgment task. As noted in Gupton and Méndez (2013) and Kitagawa and Fodor (2006), audio stimulus can control for alteration of prosody the speaker may impart when reading an example. This interpretation of prosody could change the clausal structure that the speaker (or listener) is reading into the phrase.

The format of data collection of this study provides an expansion of current understanding of this speech construction. The estar que construction, now attested in a larger geographic area, specifically in southern Ecuador and Trujillo, Peru, appears to be a structure in flux, based on its variable acceptance in terms of habituality, negation and clitic placement. Language in flux generally demonstrates varied use (Torres Cacoullos, 2012). The use of the estar que construction, judged by acceptability questionnaires, proves to be much more varied than previously attested. The hope for this study is that it has drawn attention to this unique phenomenon in Spanish and propels more research into its use and acceptance.