Received: 21-09-2024

Accepted: 22-05-2025
Published: 14-07-2025

Predicative possession choice in

Argentinian Spanish: An experimental
study
Martín Fuchs
Freie Universität Berlin
[Link]@[Link]

Kendra V. Dickinson
Rutgers University
[Link]@[Link]

Scott A. Schwenter
The Ohio State University
schwenter.1@[Link]

How to cite: Fuchs, Martín, Dickinson, Kendra & Scott Schwenter. 2025. Predicative possession
choice in Argentinian Spanish: An experimental study. RLLT 25, eds. Adam McBride, James Law
& Willis Fails. Special Issue of Isogloss. Open Journal of Romance Linguistics 11(5)/6, 1-17.
DOI: [Link]

Abstract

This study investigates the expression of predicative possession in Argentinian

Spanish, focusing on the alternation between two constructions: tener (‘have’) + NP
and estar con (‘be with’) + NP. Building on previous research, we explore the factors
that determine speakers’ choices between these constructions, particularly the
influence of temporal context and the presence of adverbs. Using a forced-choice
experimental design, participants were presented with vignettes varying in temporal
duration (durative vs. non-durative) and adverbial modification (extending, limiting,
or none). Results show a strong overall preference for the tener + NP construction,
particularly in durative contexts. However, the estar con + NP construction is more
likely to be selected in non-durative contexts, especially when a limiting adverb is
present. These findings suggest that the distinction between the two constructions is
not purely semantic but also pragmatically modulated by temporal and contextual
factors. We argue that this pattern reflects a subset-superset relationship between the
two constructions, where tener + NP can cover a broader temporal scope than estar
 2 Isogloss 2025, 11(5)/6 Fuchs, Dickinson & Schwenter

con + NP. This overlap mirrors cross-linguistic findings on possessive constructions

and aspectual distinctions, with implications for understanding grammaticalization
processes in Romance languages.

Keywords: possession, Spanish, aspect, grammaticalization.

1. Introduction
1.1. Different ways of expressing possessive relations

Spanish displays a wide array of morphosyntactic strategies to encode possessive

relations between a possessor and a possessum. Among them, the main grammatical
devices1 include attributive possession, as in (1), and predicative possession, as in (2):

(1) Spanish
El auto de María es rojo.
the car of Mary [Link] red
‘Mary’s car is red’

(2) Spanish
María tiene un auto rojo.
Mary [Link] a car red
‘Mary has a red car’

In this paper, we focus on the latter type—predicative possession—and two of

its different means of expression in the Spanish language: tener ‘have’ + Noun Phrase
(NP) and estar con ‘be with’ + Noun Phrase (NP).
From a crosslinguistic perspective, Stassen (2009) has shown that languages
generally provide four ways of encoding predicative possession: three intransitive
types and one transitive type with a direct object. Among the intransitive types, the
first is a locative or existential construction that could be paraphrased as ‘at possessor,
there is possessum’. This type of device is used by languages such as Latin, Russian,
or Irish, among others. The second type of intransitive construction is what Stassen
calls the topic construction, where the possessum is construed as the grammatical
subject of the sentence. We observe this construction in Mandarin and Thai, among
other languages. Finally, the third intransitive construction is known as the WITH-
possessive construction, where the possessum is constructed as an oblique argument,
as in most Papuan languages, for example. As for the transitive construction with a
direct object—also known as the HAVE-possessive construction—we can find it in
languages such as Norwegian, Romanian, Albanian, or Wolof, among many others.

1
Spanish can also express possessive relations in many other ways. For example, the
possessor can be instantiated by a possessive determiner agreeing in person and number
features with the possessor and in number with the possessum (e.g., Mi/s auto/s ‘my car/s’,
Nuestro/s auto/s ‘our car/s’), or by a so-called strong possessive, which shares person and
number features with the possessor and gender and number features with the possessum (e.g.,
El auto mío ‘My car’, Los autos nuestros ‘Our cars’). For other less frequent ways of
expressing possession in Spanish, see for instance Rodríguez and Howe (2023) and Suárez
Palma (2024).
 Predicative possession in Argentinian Spanish Isogloss 2025, 11(5)/6 3

Stassen also mentions that these last two types—the WITH-possessive and the HAVE-
possessive—are the two most common ways of expressing predicative possession
across the languages of the world.
Nevertheless, many languages have more than one of these construction types
available to express predicative possession in what are called split predicative
possession systems (e.g., Stolz et al. 2008, Aikhenvald 2012). While some possessive
splits are much more well-known, such as the contrast between alienable vs.
inalienable possession, or the pronominal vs. genitive constructions in the domain of
attributive possession, predicative possession splits can occur across different
dimensions: (1) they can be possessor-oriented, such that the kind of possessor leads
speakers to use one or the other construction; (2) possessum-oriented, where the
grammatical form speakers use depends on properties of the possessum; or (3)
time/manner oriented, where the choice of grammatical device is driven by
considerations about the kind of relation that obtains between possessor and
possessum.

1.2. Split predicative possession in Spanish

Our goal in this paper is to explore this last kind of split predicative possession system
in Spanish since, in this language, we observe (at least) two different means of
expression for predicative possession: a HAVE-possessive and a WITH-possessive, as in
(3) and (4) respectively:

(3) Spanish
Tengo ciudadanía argentina.
[Link] citizenship Argentinian
‘I have Argentinian citizenship’

(4) Spanish
Estoy con auto.
[Link] with car
LIT: ‘I am with car’ > ‘I have a car’

In Spanish, this split in predicative possession appears to be influenced by the

grammatical aspect of the relation between possessor and possessum. Specifically, it
has been argued (e.g., Cifuentes Honrubia 2015) that the distinction between these
forms is driven by the nature of the possessive relation –whether it is viewed as
ongoing or as a completed, punctual event. According to this account, the HAVE-
possessive, instantiated in Spanish by the tener + NP construction, expresses
inalienable possession, conveying aspectual imperfectivity, since it conveys a
possessive relation between possessor and possessum that extends over time without
an explicit ending point. On the other hand, the WITH-possessive—expressed in
Spanish by the estar con + NP construction—conveys alienable possession, and is
claimed to be related to perfectivity, implicating a punctual, limited possessive
relation. This kind of distinction between means of expression in a split possessive
system—that is, a distinction based on grammatical aspect, and whether the sentence
conveys an imperfective or a perfective possessive relation—has also been claimed to
 4 Isogloss 2025, 11(5)/6 Fuchs, Dickinson & Schwenter

be present in many other languages, such as Moroccan Arabic, Lithuanian, or Turkish

(Stolz et al. 2008).
Spanish, though, seems to present a more complex distributional pattern, since
the split displays more variability in usage. On the one hand, inalienable or permanent
possession seems to be expressed exclusively with the HAVE-possessive, as can be seen
in (5) and (6) below (we use singular ‘they’ as a gender-neutral pronoun in the English
idiomatic translations):

(5) Spanish
Tiene el pelo rubio.
[Link] the hair blond
‘They have blond hair’ (→ it’s their natural color)

(6) Spanish
Está con el pelo rubio.
[Link] with the hair blond
LIT: ‘They are with blond hair’ > ‘They have blond hair’ (→ they’ve dyed it)

In the sentence in (5), the use of the HAVE-possessive construction with no prior
context leads to the interpretation that their natural hair color is blond, emphasizing
the permanent nature of the possessive relation. By contrast, in the sentence in (6)
(again, without additional context), the use of the WITH-possessive implies that their
hair has been dyed from a different color, conveying the temporariness of the
‘blondness’ of their hair. That said, the implication that the HAVE-possessive
construction conveys the permanence of the property ascribed to the possessor can be
canceled by additional contextual information, as in (7), which is felicitous in a context
where, for example, there is a person who changes their hair color every week:

(7) Spanish
Tiene el pelo rubio esta semana.
[Link] the hair blond this week
‘They have blond hair this week’ (→ they’ve dyed it)

In this case, we can observe that the use of an adverbial like esta semana ‘this
week’ limits the temporal interval within which the person’s ‘blondness’ holds, such
that the HAVE-possessive construction conveys the same interpretation as the WITH-
possessive construction; that is, a non-permanent, temporary one. Indeed, the sentence
in (7) could be easily rendered with the WITH-possessive, but also adding the temporal
adverbial specification in (7), since it is not redundant to use the purportedly temporary
WITH-possessive with an adverbial that makes this temporariness explicit. This pattern
is in line with a generalization made by Stassen (2009: 560 et seq.), where it is claimed
that across languages, HAVE-possessives can always express WITH-possessive
meanings, but that the opposite is not the case (that is, WITH-possessives should not
always be able to express HAVE-possessive meanings, e.g., for inalienable possession
in Spanish). Besides (7) above, showing how this generalization applies to Spanish in
the case of a typically-permanent possessum, where the use of a HAVE-possessive does
not necessarily convey permanent possession, we can also observe that in cases of
 Predicative possession in Argentinian Spanish Isogloss 2025, 11(5)/6 5

more typically-temporary possessums, the alternation between the HAVE-possessive

and the WITH-possessive construction is also allowed, as in (8) and (9):

(8) Spanish
Tengo plata (conmigo) ahora.
[Link] money [Link] now
‘I have money (with me) now’

(9) Spanish
Estoy con plata ahora.
[Link] with money now
LIT: ‘I am with money (on me) now’ > ‘I have money now’

We claim that this kind of distribution, which Stassen has observed for split
possession systems, is very similar to theoretical proposals about a unified semantics
for Progressive and Imperfective operators that analyze their differences in
superset/subset terms (Deo 2009, 2015). According to these accounts, general
imperfective forms, such as the Simple Present, can express all imperfective meanings,
including the progressive one, but more specific progressive forms, such as the Present
Progressive, can only convey a progressive meaning, and thus cannot express more
general imperfective ones, such as the habitual or the continuative. This distinction
creates an implicational relationship between these forms and meanings, such that all
progressives are imperfective, but not all imperfectives are (solely) progressive. For
example, when looking at present-day Spanish, we can observe that (10) below, where
a Simple Present is used, is said to be ambiguous between a progressive and a habitual
meaning, but (11), with the use of the Present Progressive, is said to only be able to
express a progressive meaning:2

(10) Spanish
Ana fum-a
Ana [Link]
‘Ana smokes / Ana is smoking’

(11) Spanish
Ana est-á fuma-ndo.
Ana [Link] smoke-PROG
‘Ana is smoking’

Now, when we look at the Spanish predicative possessive system again in

detail, both grammatical markers that express possession—the tener + NP
construction, and the estar con + NP construction—seem to run counter to Stassen’s
generalization, since both markers appear to be compatible with adverbial
modification that entails an extended, durative state of the possessive relation between

2
This is the view advanced in traditional descriptions of the alternation between the
Simple Present and the Present Progressive in Spanish (e.g., Bull 1965, de Bruyne 1995, Roca
Pons 1958, i.a.), but we are very much aware that the pattern is more complex, determined by
linguistic and situational contextual factors, and varies across different Spanish dialects (e.g.,
Fuchs et al. 2020a, b).
 6 Isogloss 2025, 11(5)/6 Fuchs, Dickinson & Schwenter

possessor and possessum (also contra ‘perfectivity’ accounts of temporariness, such

as Cifuentes Honrubia 2015). We can see this pattern in (12) and (13):3

(12) Spanish
Siempre tengo plata.
always [Link] money
‘I always have money’

(13) Spanish
Siempre estoy con plata.
always [Link] with money
LIT: ‘I am always with money (on me)’ > ‘I always have money’

These examples illustrate that both possessive constructions in Spanish can be

used with the same NP, in the presence of an adverb like siempre ‘always’ that forces
an imperfective, habitual, meaning, indicating that previous descriptions of the
Spanish split possessive system do not sufficiently account for the observed overlap
in their distribution.
Nevertheless, accounts of the Imperfective-Progressive alternation (Deo 2015,
Fuchs 2020) also include a diachronic dimension that can help in understanding the
distribution of split possessive systems like the Spanish one. In the analysis of
progressive and imperfective markers, it has been observed that progressive
constructions slowly but steadily develop into more general imperfective ones over
time, acquiring these more general imperfective readings, such as the habitual or the
continuative, in what is known as the Progressive-to-Imperfective grammaticalization
path (e.g., Bybee et al. 1994). So, if a similar distinction is at work in the expression
of predicative possession in Spanish, we can expect WITH-possessives to increasingly
develop into more general markers of possession that are allowed to express more
durative and permanent relations between possessors and possessums, just like the
HAVE-possessives usually do.
Moreover, this kind of diachronic development has already been observed in
Brazilian Portuguese (Schwenter & Dickinson 2020). In this language, the semantic
difference between the HAVE-possessive –the construction ter + NP—and the WITH-
possessive—realized by estar com + NP– has also been characterized in aspectual
terms, but with a focus on imperfectivity (Avelar 2009). Under this account, estar
com + NP establishes a relation between possessor and possessum that is transitory –
a relation that is said to hold now, at this moment. In contrast, ter + NP expresses a
relation between possessor and possessum that is taken to hold for a longer period of
time. The diachronic relation between these two markers is mentioned by Perini, who
argues that “ter + NP can also be used to express a momentary state, but this use is not
common and sounds literary or old-fashioned. Thus, the sentence Ele tem dor de
cabeça will normally be understood only as ‘He is subject to headaches’ and not as
‘He has a headache (now)’” (Perini 2002: 277).
While ‘permanent’ versus ‘transitory’ (or ‘temporary’; see Ruvalcaba 2020 for
use of this term as applied to Spanish) is a good first description of most uses of ter +

3
We present constructed examples here for reasons of space and simplicity, but this
exact pattern can be widely found in Twitter/X and corpus data.
 Predicative possession in Argentinian Spanish Isogloss 2025, 11(5)/6 7

NP in contrast to estar com + NP in Brazilian Portuguese, Dickinson and Schwenter

(2020) show that the account fails to explain the considerable overlap that is found in
the natural use between the two constructions. They argue, thus, that the distinction
between them is not semantic, but more properly pragmatic in nature, such that certain
contexts of interpretation lead to one form being preferred over the other. To make
sense of these overlapping distributions, they propose an account very similar to
existing approaches for progressives and imperfectives in Spanish (Fuchs 2020). In
particular, they argue that the temporal interval to which ter + NP refers always
includes the one that estar com + NP can refer to, in a superordinate-subordinate
relation. In a nutshell, their argument –and the experimental evidence they provide–
supports a subset-superset relation between the two Brazilian Portuguese
constructions, such that estar com + NP makes salient the reference interval at which
the possessive relation holds, while ter + NP makes salient a superinterval thereof.
Therefore, here we ask ourselves about the strength of the association of the
HAVE-possessive construction in Spanish with the permanence of the possessive
relation, and the linking of the WITH-possessive construction to the temporariness of
the relation between possessor and possessum. To that end, we construct a forced-
choice task that helps in assessing the preferences of native speakers to use one or the
other grammatical construction in specific contexts that make salient distinctions about
the length of the temporal interval at which the possessive relation holds. Our main
hypothesis is that—just as in the Brazilian Portuguese case—the distinction between
the two grammatical forms is not semantic but pragmatic in nature, such that the
encoded meaning of both constructions overlaps considerably, and their distribution
in usage is affected by additional contextual parameters. The prediction is that estar
con + NP can only be used to refer to possessive relations that hold at a given reference
interval, while tener + NP can be used not only to refer to these kinds of possessive
relations, but also to ones that hold more extensively, at a superinterval of the reference
interval. If any differences in usage patterns arise between Spanish and Brazilian
Portuguese, we argue that they will be explained by diachronic processes where one
of the two languages is further ahead in the grammaticalization and expansion of the
WITH-possessive construction.
Before continuing, we should state that our focus on Argentinian Spanish in
this study is mainly due to the goal of investigating the system of one particular variety
of Spanish, and is not based on any dialect-specific hypotheses. Nevertheless, just as
recent studies have shown that there are completely different predicative constructions
in some varieties—e.g. the transitive andar possessive in Central American Spanish
studied by Rodríguez and Howe (2023)—we do not discount the possibility that there
could be dialectal differentiation in the use of the two instantiations of the HAVE- and
WITH-possessive constructions that we analyze in this paper. Thus, we hope that future
studies will employ methodologies similar to ours here in order to uncover possible
dialect differences.
The remainder of the paper is structured as follows. Section 2 presents the
methods of the experimental study we conducted to assess the distribution of the HAVE-
possessive and the WITH-possessive construction in Argentinian Spanish. In Section 3,
the results of the study are presented, first at a descriptive level, and then with the use
of inferential statistical analysis. Section 4 discusses the results taking into account a
variationist and diachronic perspective, and Section 5 offers conclusions.
 8 Isogloss 2025, 11(5)/6 Fuchs, Dickinson & Schwenter

2. Methods
2.1. Forced choice task study

In order to assess the distribution of the HAVE-possessive construction—tener + NP—

and the WITH-possessive construction—estar con + NP—in Spanish, we designed a
forced choice task study. To create our experiment, we used a 2x2 experimental
design, where our first independent variable was context, which presented two levels
based on a temporal and aspectual distinction: durative and non-durative contexts. Our
second independent variable was whether or not the target sentence included an adverb
that made salient the duration of the interval during which the possessive relation held
between possessor and possessum. Specifically, we included extending temporal
adverbs in the case of durative contexts, and limiting temporal adverbs for non-
durative contexts. We did not include examples in which there was a clash between
the aspectual value of the context and the semantic content of the adverb. An example
of each case is given in Table 1.

Table 1. Experimental stimuli across all conditions, crossing the two independent variables:
duration of the context (durative vs. non-durative) and presence or absence of an adverb that
made salient the temporal nature of the possessive relation (limiting vs. extending vs. none).

Choice A Choice B
Duration Adverb Context (example)
(tener) (estar con)
Pedro comió hace dos horas, pero
quiere comer de nuevo. ¿Qué es Estoy con
Tengo
más probable que diga? hambre
non- hambre
none
durative
‘Pedro ate two hours ago, but he ‘I am with
‘I’m hungry’
wants to eat again. What is he hunger’
more likely to say?’
Pedro comió hace dos horas, pero
Ya estoy
quiere comer de nuevo. ¿Qué es Ya tengo
con hambre
más probable que diga? hambre
non-
limiting
durative ‘I am
‘Pedro ate two hours ago, but he ‘I’m already
already with
wants to eat again. What is he hungry’
hunger’
more likely to say?’
Pedro come mucho, pero para él
nunca es suficiente. ¿Qué es más Estoy con
Tengo
probable que diga? hambre
hambre
durative none
‘Pedro eats a lot, but it is never ‘I am with
‘I’m hungry’
enough for him. What is he more hunger’
likely to say?’
Pedro come mucho, pero para él
Siempre Siempre
nunca es suficiente. ¿Qué es más
tengo estoy con
probable que diga?
hambre hambre
durative extending
‘Pedro eats a lot, but it is never
‘I’m always ‘I am always
enough for him. What is he more
hungry’ with hunger’
likely to say?’

Each experimental item consisted of a contextual vignette aimed at establishing

the context as either durative or non-durative, as exemplified in Table 1. We created
 Predicative possession in Argentinian Spanish Isogloss 2025, 11(5)/6 9

pairs of vignettes that were maximally similar, including the same topic and
possessum, varying only in their establishment of the contextual duration, for a total
of 24 different vignettes (12 general scenarios that varied across the dimension of
duration, for a total of 24 vignettes). Possessums were of four different kinds (mental
states, illnesses, objects, and abstract concepts) following existing typologies of
possessums (see Heine 1997, Stassen 2009, Stolz et al. 2008); these are kept as a
random effect in this study, since even though we included each kind in the vignettes,
we are not able to guarantee that other possessums within each category that we did
not employ in the experimental stimuli would behave similarly.
Each of the 24 contextual vignettes was followed by two target stimuli, one
including tener + NP and one including estar con + NP. The experimental task given
to participants was to choose one of the two options (tener + NP or estar con + NP) as
the most likely continuation, given the context and the optional adverb. We utilized
each contextual vignette twice in our experiment. In one version, the target stimuli that
followed included a temporal adverb that made salient the temporal nature of the
possessive relation, and in the other, no temporal adverb was included in the target
stimuli. In this way, we created a total of 48 experimental items for the study.
With these items, we created four different versions of the experiment using a
Latin Square Design, so that each participant only saw one condition per experimental
stimulus. Consequently, each participant saw 24 items in total (12 target items and 12
fillers), in a fully randomized sequence. The experiment was hosted on Qualtrics
(Qualtrics, Provo, UT), and distributed via social media and researcher contacts.

2.2. Participants

Our participants consisted of 59 Spanish-native speakers (45 female) from Argentina,

for a mostly homogenous sample. Almost all of them (n = 56) were from Buenos Aires,
with the exceptions of one speaker from Santa Fe, one from Córdoba, and one from
Misiones. They were between 29 and 69 years old (mean = 44.2 years old), had all at
least completed high school (53 of them had a college degree or higher), and all self-
identified as middle or upper-middle class. We collected a total of 707 data points for
our analysis.

2.3. Statistical Methods

Data were analyzed using R software (R Core Team 2024). We performed initial
descriptive analysis of the data and used the ggplot2 package (Wickham 2016) for
visualization. Then, we conducted an inferential analysis using the lme4 package
(Bates et al. 2015). We constructed a series of nested mixed-effects logistic regression
models with possessive type as the dependent variable, and context, temporal adverb,
and participant demographic characteristics as potential predictors. For our linguistic
predictors, we utilized treatment contrast coding due to the experimental design of our
study. In addition to fixed effects, all nested models included random intercepts for
respondent and possessum. The former was included to account for repeated measures,
and the latter to account for well-established differences in possessive type selection
according to possessum (Stolz et al. 2008), which, while of interest, are not central to
the present analysis. We then compared our models using the anova function (Fox &
Weisberg 2018), a likelihood ratio test, to determine the optimal model for the data.
 10 Isogloss 2025, 11(5)/6 Fuchs, Dickinson & Schwenter

Finally, we used the party package (Hothorn et al. 2006) to generate a conditional
inference tree (CIT), which uses regression and p-values to identify the best splits in
predictors, in order to visualize possible interactions in our data.

3. Results
3.1. Description of the data

Data for the analysis consist of the aforementioned 707 data points from the
experimental forced-choice task. Overall results show that tener + NP was selected by
respondents more frequently overall (74.4%, n=526) than was the estar con + NP
construction (25.6%, n=181). Nevertheless, respondent possessive type preference
showed considerable variation according to both contextual and grammatical factors.
The first notable pattern in our data, aside from the overall preference for
tener + NP, is the even greater selection of this construction by respondents in durative
contexts (79%, n=279) compared to non-durative contexts (69.8%, n=247).
Conversely, respondents were more likely to select estar con + NP in non-durative
(30.2%, n= 107) than in durative contexts (21%, n=74). This pattern in our data shows
that durative contexts lead to increased selection of tener + NP, and therefore
decreased selection of estar con + NP. This difference is visualized in Figure 1.

Figure 1. Distribution of possessive selection by context

The second notable pattern in our data relates to the differences in rates of
possessive type selection according to the presence or absence of temporal adverbs.
The presence of a limiting adverb increased the likelihood of respondent selection of
estar con + NP (35.6%, n=63) compared to extending adverbs (24.3%, n=43) and
adverb absence (21.9%, n=75). The full distribution of possessive selection according
to content and temporal adverb is included in Table 2. These differences are also
 Predicative possession in Argentinian Spanish Isogloss 2025, 11(5)/6 11

visualized in Figure 2, which highlights the role of temporal adverbs in modulating

possessive construction selection.

Figure 2. Distribution of possessive selection by temporal adverb

Table 2. Rates of possessive type selection according to content and adverb

tener + NP estar con + NP

Context Adverb
(74.4%, n=526) (25.6%, n=181)
82.4% 17.6%
none
+ durative (n=145) (n=31)
(n=353) 75.7% 24.3%
extending
(n=134) (n=43)
75.1% 24.9%
none
- durative (n=133) (n=44)
(n=354) 64.4% 35.6%
limiting
(n=114) (n=63)

Overall, the distribution of possessive type selection among our participants

shows disparate patterns according to context and the presence/absence of a temporal
adverb. To further examine these potential effects, we now turn to the inferential
analysis of the data.

3.2. Inferential analysis

The best-fit model for these data included a significant interaction between context
and adverb, with respondent and possessum as random intercepts. The output of this
model is provided in Table 3, and shows the comparative likelihood of participants
 12 Isogloss 2025, 11(5)/6 Fuchs, Dickinson & Schwenter

selecting tener + NP compared to estar con + NP in non-durative contexts with no

temporal adverb versus other factor combinations.

Table 3. Output of best-fit mixed-effects logistic regression model

Estimate Std. Error z value p value

(Intercept) 1.39223 0.38726 3.595 0.000324

+ durative Extending adverb 0.54034 0.29156 1.853 0.063843
- durative No adverb 0.05727 0.27592 0.208 0.835564
- durative Limiting adverb -0.66463 0.26201 -2.537 0.011193*

First, this output shows that in durative contexts, the presence of an extending
adverb does not significantly increase the likelihood of selection of tener + NP among
our participants. We interpret this result as reflecting the default status of tener + NP
in the predicative possessive paradigm in present-day Argentinian Spanish. Similarly,
non-durative contexts with no temporal adverb do not demonstrate meaningful
differences in use of tener + NP compared to durative contexts, again supporting the
view of tener + NP as the default possessive. Nevertheless, this output shows that the
presence of a limiting adverb in non-durative contexts significantly decreased the
likelihood of participant selection of tener + NP, and correspondingly increased the
likelihood of participant selection of estar con + NP, effectively illustrating its locus
in the space of predicative possession tested. Figure 3 visualizes this interaction.
As can be observed in Figure 3, tener + NP was selected as the predicative
possessive type of preference by participants across all contexts, thereby corroborating
its expansive range across the realm of the predicative possessive space. Nevertheless,
participants showed an even greater preference for tener + NP in durative contexts
(Node 5), independent of the presence or absence of a temporal adverb compared to
non-durative contexts. This overarching preference for tener + NP, as well as the lack
of impact of temporal adverbs in durative contexts on selection of this construction,
highlight its broader applicability in comparison to estar con + NP. By contrast, while
non-durative contexts without a co-occurring temporal adverb increase the likelihood
of participant selection of estar con + NP compared to non-durative contexts (Node
4), the greatest rates of selection of estar con + NP in non-durative contexts are found
in the presence of a limiting adverb (Node 3). Overall, these findings highlight the role
of tener + NP as the less contextually-restricted variant as well as the role of limiting
adverbs in providing contextual support for the use of the less-frequent estar con + NP
construction in non-durative contexts.
 Predicative possession in Argentinian Spanish Isogloss 2025, 11(5)/6 13

Figure 3. Conditional Inference tree of interaction between context and adverb

4. Discussion

Overall, these results show that respondents variably select both possessive forms in
Argentinian Spanish across all experimental conditions. However, they choose tener +
NP—the HAVE-possessive construction– more often, and especially in durative
contexts, while they decide to use estar con + NP—the WITH-possessive construction–
in non-durative contexts, particularly when the temporariness of the possessive
relation is made salient by the presence of a limiting adverb. Though not unexpected,
these results provide empirical corroboration for the view that tener + NP is the default
predicative possessive construction in this variety and most likely across Spanish
dialects more generally. Most importantly, and in line with our predictions, our
findings show that the meaning of these possessive constructions can be modulated by
both grammatical devices (such as temporal adverbs) and discourse conditions (such
as the aspectual nature of the situational context), which function as modifiers of the
temporal and aspectual nature of the possessive relation between possessor and
possessum.
We have therefore provided empirical verification of the “splitness” (e.g., Stolz
et al. 2008) of the predicative possessive system in Argentinian Spanish, but at the
same time we have highlighted the overlapping nature of the two grammatical forms
included in our study that help make up this system. We have done this by changing
the way that the aspectual distinction between these forms is conceptualized. Instead
of an analysis that distinguishes the HAVE-possessive construction and the WITH-
possessive construction in terms of imperfectivity vs. perfectivity, we claim that these
results are better explained by sustaining that both tener + NP and estar con + NP
convey imperfectivity. Their true aspectual difference concerns the temporal interval
 14 Isogloss 2025, 11(5)/6 Fuchs, Dickinson & Schwenter

that each construction refers to: in the case of estar con + NP, the interval at which the
possessive relation holds is co-extensive with a contextually determined reference
interval; on the other hand, in the case of tener + NP, the possessive interval is co-
extensive with a superinterval of the reference interval determined by the context. This
subset/superset relation is what allows for a broader usage of the HAVE-possessive
construction—i.e., tener + NP—over the WITH-possessive construction—i.e., estar
con + NP. This pattern also corresponds to the well-known use of tener + NP in
Spanish for inalienable possession, as in Tengo dos manos ‘I have two hands’ (where
estar con + NP is regularly not used4), while still being compatible with alienable
possession contexts.
When we take into consideration a cross-Romance diachronic dimension to
these results, we observe differences between Argentinian Spanish and Brazilian
Portuguese (Schwenter & Dickinson 2020). When looking at Spanish, we see that
estar con + NP appears to signal the temporariness of the possessive relation by
making salient the (reference) interval at which the relation between the possessor and
the possessum holds. To be felicitous, this marker requires both situational context
support—in the form of a non-durative context—and linguistic material support—
such as the presence of a limiting temporal adverb. A following stage in the potential
generalization pathway of this marker can be found in present-day Brazilian
Portuguese, where Schwenter and Dickinson (2020) have shown that the
corresponding WITH-possessive construction—estar com + NP—is able to appear in a
broader set of non-durative contexts, regardless of the presence/absence of a limiting
temporal adverb. In addition, Schwenter and Dickinson (2020) found, using similar
experimental materials, that the estar com + NP construction was actually more
frequent in the Brazilian Portuguese data than the competing ter + NP construction, a
result not found in the Argentinian Spanish data presented here.5 The expectation is
that the estar com + NP construction will later generalize as well to durative contexts,

4
Both reviewers ask whether estar con is ruled out with inalienable possession (and
relatedly, with kinship terms—Tengo una hermana ‘I have a sister’, but not Estoy con una
hermana ‘LIT: I am with a sister’) on semantic or pragmatic grounds. Under this account, we
argue that the incompatibility is pragmatic; there are not very few contexts where a proposition
involving inalienable possession (or a kinship term) would hold for a specific reference
interval without holding for a superinterval thereof. That said, one could come up with
contexts that allowed for the use of estar con and inalienable possession: for example,
someone who has lost an arm in an accident and uses a prosthetic one when they are in public
could say Estoy con dos brazos ‘LIT: I am with two arms’ when they are using the prosthetic
arm, and Estoy con un brazo ‘LIT: I am with one arm’, when they are not using it. We did not
include these kinds of items in our experimental task since they are not easily amenable to the
kind of contextual manipulation that we were interested in, and are very uncommon in natural
spoken data.
5
There are other, non-quantitative, sources of difference between Argentinian Spanish
and Brazilian Portuguese, e.g. the availability of a cleft-like exclamative that can only be
rendered with tener + NP in Spanish (¡Qué hambre que tengo! ‘How hungry I am!’) that can
occur with both the HAVE-possessive and with the WITH-possessive in Brazilian Portuguese
(Que fome que tenho! Que fome que estou/tô!). This latter version is not possible in Spanish
(*¡Qué hambre que estoy!). We take this dual possibility in Brazilian Portuguese to be another
piece of evidence that the WITH-possessive is more generalized and can thus be used in a
broader set of contexts than in Argentinian Spanish.
 Predicative possession in Argentinian Spanish Isogloss 2025, 11(5)/6 15

thereby constraining the use of the now contextually more general HAVE-possessive
construction.

5. Conclusions

In this paper, we set out to understand the distribution of partially overlapping forms
for the expression of predicative possession in Argentinian Spanish; namely a HAVE-
possessive construction—the tener ‘have’ + NP combination—and a WITH-possessive
construction—made up of estar con (infinitive ‘be’ and ‘with’) plus an NP. The results
of the experimental task carried out with native Argentinian speakers showed that
these two constructions are not equivalent in usage, but that their encoded semantics
is equal, and the distinction in their distributions arises from pragmatic parameters that
manipulate the length of the temporal interval at which the relation between the
possessor and the possessum holds. The WITH-possessive construction can be used to
make reference to a contextually salient reference interval that overlaps with the
possessive relation, while the HAVE-possessive construction can also refer to a
superinterval of that relation, but does not do so necessarily.
In line with previous research in Brazilian Portuguese (Schwenter & Dickinson
2020), we found that “permanent” against “temporary” is a good first description for
most uses of tener + NP and estar con + NP in Argentinian Spanish, but this account
implies that the two forms are in complementary distribution and fails to explain the
considerable overlap that is found in usage between the two grammatical devices. The
distinction between the two constructions is better explained by considering the
temporal interval that each of these constructions makes salient, as stated above. In a
nutshell, tener + NP always includes the reference interval of estar con + NP, showing
a strong parallelism between these competing possessive forms and imperfective and
progressive operators, whose interpretation is also contextually dependent along the
same parameters (Deo 2009, 2015; Fuchs et al. 2020a, b).
All in all, we have shown that a phenomenon in Spanish that has long been
thought to have a purely semantic or grammatical basis has a crucial additional
pragmatic dimension, with specific contexts triggering different interpretations of
overlapping linguistic expressions. In future research we hope to extend this approach
to other forms found in some Spanish varieties that likewise appear to overlap with the
HAVE-possessive construction and with the WITH-possessive construction examined
here. In particular, we are interested in the use of traer ‘to bring’ + NP (see Rábago
Tanori & Melis 2005, Ruvalcaba 2020) in Mexican Spanish and its overlap with
tener + NP, and in the use of andar ‘to walk, go around’ con + NP in both Mexican
and Central American Spanish varieties (Rodríguez & Howe 2023), which competes
with estar con + NP, as well as in the recently-described innovative direct transitive
use of andar + NP in Central American varieties (ibid.).

Acknowledgments

We would like to thank the audience at LSRL54, as well as the editors of this volume
and two anonymous reviewers for their helpful and constructive comments. All errors
and omissions remain our own.
 16 Isogloss 2025, 11(5)/6 Fuchs, Dickinson & Schwenter

References

Aikhenvald, Alexandra Y. 2012. Possession and ownership: a crosslinguistic

perspective. In A. Y. Aikehenvald, & R. M. W. Dixon (eds), Possession and
ownership: A cross-linguistic typology, 1-64. Oxford: Oxford University Press.

Bates, Douglas, Maechler, Martin, Bolker, Ben, & Steve Walker. 2015. Fitting linear
mixed-effects models Using lme4. Journal of Statistical Software 67(1): 1-48.
[Link]

Bull, William E. 1942. New Principles for Some Spanish Equivalents of "To Be".
Hispania 25(4): 433-443. [Link]

Cifuentes Honrubia, José Luis. 2015. Construcciones posesivas en español. Leiden:

Brill/Rodopi.

de Bruyne, Jacques. 1995. A Comprehensive Spanish Grammar. Oxford: Blackwell.

Deo, Ashwini. 2009. Unifying the imperfective and the progressive: Partitions as
quantificational domains. Linguistics and Philosophy 32: 475-521.
[Link]

Deo, Ashwini. 2015. Diachronic semantics. Annual Review of Linguistics 1: 179-197.

[Link]

Fox, John, & Sanford Weisberg. 2018. An R companion to applied regression.

California: Sage Publications.

Fuchs, Martín. 2020. On the Synchrony and Diachrony of the Spanish Imperfective
Domain: Contextual Modulation and Semantic Change. PhD Dissertation, Yale
University.

Fuchs, Martín, Deo, Ashwini, & María Mercedes Piñango. 2020a. The progressive-to
imperfective shift: Contextually determined variation in Rioplatense, Iberian, and
Mexican Altiplano Spanish. In A. Morales-Front, M. J. Ferreira, R. P. Leow, & C.
Sanz (eds.), Hispanic linguistics: Current issues and new directions, 119–136.
Amsterdam: Johns Benjamins. [Link]

Fuchs, Martín, Piñango, María Mercedes, & Ashwini Deo. 2020b. Operationalizing
the role of context in language variation: The role of perspective alignment in the
Spanish Imperfective domain. In S. Löbner, T. Gamerschlag, T. Kalenscher, M.
Schrenk & H. Zeevat (eds.), Concepts, Frames and Cascades in Semantics, Cognition
and Ontology, 201-217. Berlin: Springer. [Link]
3_10

Heine, Bernd. 1997. Possession. Cambridge: Cambridge University Press.

 Predicative possession in Argentinian Spanish Isogloss 2025, 11(5)/6 17

Perini, Mário. 2002. Modern Portuguese: A reference grammar. New Haven: Yale
University Press.

R Core Team. 2024. R: A language and environment for statistical computing. R

Foundation for Statistical Computing.

Rábago Tánori, Álvaro, & Chantal Melis. 2005. El sentido posesivo de traer.
Lingüística Mexicana 11(2): 227-247.

Roca Pons, José. 1958. Estudios sobre perífrasis verbales del español. Madrid: CSIC.

Rodríguez, Shannon P., & Chad Howe. 2023. Big data and small dialects: Transitive
andar in Central American Spanish. In B. Baird, O. Balam, & M. Parafita Couto (eds.),
Linguistic Advances in Central American Spanish, 168-192. Leiden: Brill.
[Link]

Ruvalcaba, Christian. 2020. “¿Qué traes güey?”: An analysis of Spanish clausal

possession. In J. J. Colomina Almiñana, & S. Sessarego (eds.), Language Patterns in
Spanish and Beyond, 52-74. London: Routledge.
[Link]

Schwenter, Scott A., & Kendra V. Dickinson. 2020. A distinct aspectual analysis of
predicative possession in Brazilian Portuguese. Proceedings of the Linguistic Society
of America 5(1): 242-256. [Link]

Stassen, Leon. 2009. Predicative possession. Oxford: Oxford University Press.

Stolz, Thomas, Kettler, Sonia, Stroh, Cornelia, & Aina Urdze. 2008. Split possession:
An areal-linguistic study of the alienability correlation and related phenomena in the
languages of Europe. Amsterdam: John Benjamins.

Suárez Palma, Imanol. 2024. Inalienable possession (and lack thereof) in Spanish.
Studies in Hispanic and Lusophone Linguistics 17(1): 109-136.
[Link]

Wickham, H. 2016. ggplot2: Elegant graphics for data analysis (R package version
3.3.5). Springer. [Link]