The Status of the Extension of estar in Cuban Spanish

Gabriela G. Alfaraz
Michigan State University
Abstract
This paper presents variationist sociolinguistics research on the copula estar with
predicate adjectives in Cuban Spanish, a variety in which it appears to have gone
largely uninvestigated. To examine its social and linguistic distribution, a real-time
study with data from the 1960s and 1990s was coupled with an apparent-time study
with data from the 1990s. Findings showed that generation and adjective type were
significant factors constraining the variation. The comparison of different
generations in real and apparent time suggested that the frequency of estar had
increased significantly in the younger generation compared to older ones, and it had
remained stable for the two age cohorts studied in real time. Results for following
adjective showed that the extension of estar was favored in two of eight adjective
classes. These findings suggest that Cuban Spanish has experienced a change over
time in the frequency and distribution of innovative estar with predicate adjectives.
1.

Introduction

This paper presents research on the variation of the Spanish copulas ser and estar in
Cuban Spanish. It provides an account of the variation of the copulas with predicate
adjectives in a variety in which this phenomenon appears not to have been studied.
As this research is carried out within the variationist tradition in sociolinguistics, it
contributes to previous variationist work that has examined linguistic and social
factors, i.e. type of adjective, style, age, gender, socioeconomic status (CortsTorres 2004, Gutirrez 1992, 1994, de Jonge 1993, Ortiz Lpez 2000, Salazar 2007,
Silva-Corvaln 1986, 1994), in order to identify those that condition the variation
and to estimate their degree of influence. In variationist sociolinguistics, the
description of the linguistic and social distribution of variable linguistic forms is
interpreted within the larger view of the variation as either stable or a change in
progress. The primary objective of the research reported on here is to explore
whether Cuban Spanish is undergoing a change in the use of estar as has been
reported for other varieties of Spanish (Gutirrez 1992, 1994, Silva-Corvaln 1986,
1994). Linguistic and social factors are examined to describe and compare the
variation in both real and apparent time in order to explore change over time.

Studies in Hispanic and Lusophone Linguistics

Volume 5, Issue 1
Spring 2012

Gabriela G. Alfaraz

It is well known that the Spanish copulas have been involved in change over
the course of many centuries. The evolution of the copulas has involved the
expansion of estar and the restriction of ser (Andrade 1919, Pountain 1982, SilvaCorvaln 1986, 1994, Va-Cerd 1982). After examining the acquisition of the
copulas in children, Sera (1992) noted that the expansion of estar could eventually
erase distinctions between ser and estar in contexts in which either may occur.
Their synchronic status was explored in a seminal sociolinguistic study carried out
by Silva-Corvaln (1986, 1994) with Mexican-origin subjects in Los Angeles that
convincingly showed that there has been a change in the distribution of estar and
that its expansion to ser contexts can be observed in some present-day varieties. In
reference to the change, Silva-Corvaln wrote: the extension to new contexts
represents a more advanced stage in a continuous process of syntactic-semantic
extension of the copula estar throughout the history of Spanish (p.93). SilvaCorvaln used the term extension of estar for the process of change and referred to
it using the terminology common in variationist sociolinguistics as an innovation or
innovative. Similarly, in the present paper, extension of estar and innovative estar
are used.
1.1 Background
At the center of investigations examining linguistic change from a variationist
framework is the idea that linguistic variation is systematic and constrained by
linguistic and social factors (Labov 1963). The study of variation has been based on
the notion that the variants being studied are semantically equivalent, or two ways
of saying the same thing (Labov 1972, 1978), and while it is well known that for
phonological variables this definition is unproblematic, much has been written
about the equivalence of variants when higher level variables are studied. In regards
to variables beyond the phonological level, Buchstaller (2009, p.1015) noted: some
studies do not problematise the issue and refrain from explicitly defining a variable
[...]. Others define a sociolinguistic variable but circumvent the question of its
semantic load and focus instead on the fact that the variants are united by a common
(mainly functional) denominator. Sociolinguistic studies of the variation of the
copulas ser and estar before adjectives, particularly the extension of estar, have
taken the latter approach, studying the variation without problematizing the
semantic equivalence of the variants.
The copulas ser and estar can both appear before predicate adjectives without
being variants of one variable, as shown in (1), in which ser (a) and estar (b) have
different meanings: (a) indicates that Julio is old in terms of chronological age and
(b) conveys the meaning that Julio appears or looks old but he is not necessarily
chronologically old. Yet, the copula estar in (b) could be an innovative use, and in
this sense, both (a) and (b) mean that Julio is old.

The Status of the Extension of estar in Cuban Spanish

(1) a.
b.

Julio es viejo.
Julio est viejo.

Some background of the long debate on the copulas is in order, given that
linguists have long grappled with their notorious behavior preceding predicate
adjectives (Bolinger 1947, Bull 1942, Clements 1988, Crespo 1946, Lujn 1981,
Falk 1979, Franco & Steinmetz 1983, 1986, de Mello 1979, Navas-Ruiz 1963,
Roldn 1974). On the one hand, grammars and textbooks have long explained the
distribution of ser and estar with predicate adjectives as expressing relative
duration, where ser conveys permanence and estar temporariness (Bello 1847).
Although this description can be productive, counterexamples abound (Franco &
Steinmetz 1983), which led Andrade (1919, p.19) to write: What is the ratio of
permanence between wealth and youth? [...]. Is laziness less permanent than
honesty? [...].To take permanent in the sense of changeless is utterly impossible, as
very few of all the possible properties and conditions predicated by ser are exempt
from change. Bull (1942) similarly pointed out problems with the interpretation of
the copulas as conveying permanent or inherent and temporary or accidental and
advanced the theory, based on Alonso & Henrquez Urea (1939), that the defining
characteristic of estar was not that it was temporary or accidental, but that a change
had already occurred, in which case, the new state could be permanent.
The extension of estar was evident in some of the explanations and examples of
the use of the copulas in early discussions. For instance, differences between the
copulas were blurred in Crespos (1946) analysis when he argued that estar could
refer to permanent, inherent, or normal states in addition to temporary and
accidental ones. Crespo expanded the theory that ser referred to the subjects
membership in a class and estar to its condition or state, arguing that estar not only
referred to a state or condition of the subject established through a comparison to
itself, but that it also could refer to a comparison of the subject with other similar
objects, as for example when a tourist contemplates the ceiling of the Sistine Chapel
for the first time and utters qu hermoso est esto how beautiful this is (p.49).
Bolinger (1947) shared Crespos view, arguing that in esta carne es buena this
meat is good, ser is used for the norm and compares the meat to others in the same
class, and that in esta carne est buena this meat is good the comparison is also
with items in the same class and obtains a departure from the class norm. The
important point is that for both copulas, the reference is to a class. Bolinger also
added the idea of comparison with an archetype: estar is used for comparisons
within a given genus: comparisons of a thing with its archetype or with previous or
succeeding states of itself (p.365). Franco & Steinmetz (1983, 1986), although
their primary claim was that ser established a reference to the real world, a
comparison of x to y, and estar a comparison of the entity to itself, of x to x, they
added that if the entity was new to the speaker, estar compared it with an imagined
or anticipated one. Thus, research on the copulas appears to have been complicated

Gabriela G. Alfaraz

by the intrusion of the extension of estar. As noted earlier, the few sociolinguistic
studies carried out on the copulas have based their analysis on the assumption that
they are variants of one variable preceding predicate adjectives.
1.2 Variationist studies
As noted above, Silva-Corvalns (1986, 1994) seminal study examined ser and
estar as linguistic variants to investigate the internal factors relevant to the variation
and the social factors that could influence internal linguistic changes. She found that
Mexican-origin subjects in Los Angeles used estar as well as ser in descriptions of
individuals who were unknown to their interlocutor, as shown in (2), in which the
size adjectives grande and chico occur with both copulas; estar is used with grande
big to describe the size of a persons nose. Silva-Corvaln found that in
conversational interviews the extension of estar occurred at a rate of 34%.
(2) Mira, la nariz de ella no es como la ma; est un poco grande, pero- anchita.
Ah, sus ojos son chicos como los mos. Su cara, t sabes, es- bueno, era muy
bonita mi madre.
Well, her nose is not like mine; its a little big, but- wide. Ah, her eyes are
small like mine. Her face, you know, its- well, she was very pretty, my
mother. (Silva Corvaln 1994, pp.101-102)
To probe whether innovative estar was used because speakers assumed shared
knowledge of the entity with the interlocutor, Silva-Corvaln gave a fill in the blank
questionnaire to a target group of Mexican Spanish speakers and two control
groups, one of Chileans and another of graduate students from different regions.
The results substantiated the findings of the conversational data. The rate of estar
for Mexican Spanish speakers exceeded the frequency of the other two groups: for
the target group it was 14.8%, whereas for the control groups, it was 4.2% for the
Chileans, and 1.5% for the graduate students. Silva-Corvaln, thus, concluded that
there was indeed variation of ser and estar with adjectives and that the use of estar
was not necessarily triggered by the discourse situation and presupposed
background familiarity with an entity, but rather by a change in the copula estar,
which was perhaps accelerated in the language contact situation.
When Silva-Corvaln examined the types of adjectives used with estar, she
found that it occurred with an overwhelming number of adjectives of age (78%),
then size (53%), and sensory characteristics, such as taste, smell, and hearing,
(47%); on the other hand, the adjectives with the lowest rates of estar were those of
moral value for animate subjects, such as honesto and respetable (13%), color
adjectives (11%), and adjectives referring to the social categories of animate
subjects, pobre poor and mexicano Mexican (4%).

The Status of the Extension of estar in Cuban Spanish

Gutirrez (1992, 1994) replicated Silva-Corvalns study with native speakers

in Morelia, Michoacn, Mexico, a speech community removed from language
contact influence, to explore questions that arose from Silva-Corvalns work. The
results of this study revealed generational differences that suggested a change in
progress in this speech community, too. Gutirrez found that innovative estar
occurred at an overall rate of 18%, with 16% in conversational interviews when
individuals described an entity unknown to the interlocutor. The frequencies of
innovative estar according to type of adjective were lower than Silva-Corvalns,
but the distribution was generally similar: the highest rates were with adjectives of
description (44%) and age (43%), then size (34%) and physical appearance (33%),
while low rates were found for adjectives that expressed evaluations (12%), moral
value (2%), and social status (1%). The example in (3) shows the extension of estar
with the size adjective grande big.
(3) ...como el camin era muy grande tena asiento de sobra, ... no ... como el
carro de mi hijo tambin est muy grande, ... tiene metro y medio ms grande
que todos
... since the bus was (ser) very big it had lots of seats ..., ... like my sons car
that is (estar) also very big, it has a meter and a half more than the others ...
(Gutirrez 2003, p.172)
In a subsequent study on Mexican Spanish, Corts-Torres (2004) found 23%
innovative estar overall in the city of Cuernavaca, located just south of Mexico
City, in her study of 36 men and women. The rates for conversations and a
questionnaire were considerably different, higher in the questionnaire (31%) than in
conversational data (9%). The results of a multivariate analysis with Goldvarb
showed that type of adjective was the most important factor, followed by style
(conversation vs. questionnaire) and then educational level. The results for type of
adjective showed that adjectives of physical appearance (feo ugly) and age (nuevo
new) favored innovative estar, adjectives of size and evaluation (bueno good)
were neutral, whereas those for personal characteristics (amable kind) strongly
disfavored the variant. The findings for educational level are consistent with what is
expected for nonstandard variants: rates of innovative estar increase as educational
level decreases. Gender was not a significant factor; men and women each had
frequencies of 50%. Three age groups were studied, distributed among the ages of
18-30, 32-49, and 50+. An interesting finding was that age was not significant,
which suggested stable variation rather than a change in progress, as was found in
the speech community studied by Gutirrez (1992, 1994). In Cuernavaca, the
frequency of innovative estar was higher in the middle age group (46%) than in the
youngest (34%) or the oldest (20%), but these age differences were not significant
when all factors were considered together.

Gabriela G. Alfaraz

In a study aimed at investigating the progression of linguistic change, which

does not advance at a regular rate and may remain stable over many centuries
without going to completion (Labov 2001), de Jonge (1993) examined estar from
the perspective of its having remained relatively stable during the last several
hundred years. De Jonge limited the study to adjectives of age in data from Mexico
City and Caracas, Venezuela, in which the increase of innovative estar had
remained stable for the last one to two decades. De Jonge argued that the two
varieties were at different stages in the change, after finding a statistically
significant difference in the frequency of innovative estar (36% Mexico City vs.
53% Caracas). In terms of the primary goal of examining the advancement of estar,
de Jonge concluded that it had remained stable over the span of time studied.
Salazars (2007) study of the copulas in New Mexico Spanish did not reveal
significant results for age that pointed to a change in progress in this variety among
20 subjects. Level of education, however, was found to contribute significantly to
the variation, with higher educational levels disfavoring the extension of estar.
Although gender was not significant, the frequency of innovative estar was higher
among males (54%) than females (46%). As in previous studies, type of adjective
was found to have the greatest influence on the occurrence of the variant. Adjectives
of physical description and states of being favored estar, as did adjectives of age.
Descriptive and evaluative adjectives both disfavored innovative estar, but the
effect of the latter was considerably stronger.
Evidence from variationist studies on the copulas have revealed some situations
in which generational differences suggest a change in progress (Gutirrez 1992,
Silva-Corvaln 1994) and others in which the lack of significance of age points to
stable variation (Corts-Torres 2004, de Jonge 1993, Salazar 2007). The current
research will add to these studies by examining the social and linguistic distribution
of the extension of estar to determine its status as stable variation or one in which
estar is progressing in adjective contexts as part of a process of change.
2.

Study design

To examine the extension of estar with predicate adjectives, this study compared
generational groups in both real and apparent time. The real time approach provides
a look at the frequency of the variants at different points in time with recordings
made thirty years apart, in the 1960s and the 1990s. Studies in real time can reveal
information about the status of a variable, whether it is involved in change or stable,
in a speech community. Studies in real time can either compare the same group of
speakers at various points in time or they can use a sample made up of different
speakers as representative. Previous studies can be used to establish a basis of
comparison or real-time comparisons can be carried out with data recorded earlier,
which has proven to be a valuable resource when studying linguistic variation
(Bailey, Maynor & Cukor-Avila 1991). Apparent time studies are used to

The Status of the Extension of estar in Cuban Spanish

reconstruct time differences to explore linguistic changes (Cedergren 1973, Labov

1966) when data from real-time studies are not available. In apparent time studies,
variants are studied across different age groups at one point in time; changes in
progress are revealed as differences in the distribution of variants across age groups,
with the younger tending to have increased rates when there is change from below.
For the real-time analysis, this study used data recorded in 1968-69 and 199698. The subjects were native speakers of Spanish who had recently arrived from
Cuba. The data are from conversations and informal interviews, and the two sets of
data are comparable, as they involved recent immigrants with similar demographic
profiles recorded in similar interview situations. The 1990s recordings were made in
Miami; subjects were interviewed in their homes, most often in pairs, and the
recordings made in the 1960s are a subset of the Cuban data recorded for a
dialectology study on the major cities of the Spanish-speaking countries of Latin
America and the Iberian Peninsula (Proyecto Sobre el Estudio del Espaol en
Ciudades Iberoamericanas). The Cuban part of the study was carried out in Miami,
with subjects recently arrived from Cuba, interviewed in pairs in their homes or at
Miami Senior High, where they attended evening English classes.
The generational groups and the time they were recorded are shown in Table 1.
Speakers from the 1960s corpus, who were, based on their self-descriptions,
between the ages of 30-50 at the time, are compared with a group of speakers from
the 1990s, between the ages of 30-43. A third group of speakers who were between
the ages of 62-77 in the late 1990s were also examined in order to compare them to
their age cohorts recorded in the 1960s. Data from 50 subjects were analyzed, 24
from the 1960s corpus, and 26 from the 1990s corpus, 20 younger and 6 older
speakers.
Time
Age

1960s
30-50
-

1990s
62-77
30-43

Table 1. Groups according to generation and time

Adjectives were classified according to the types outlined in Dixon (2004), in
order to provide a replicable classification. Adjective types included dimension,
age, value, color, physical property, human propensity, and speed. Examples of the
adjectives from Dixon and examples from the Cuban Spanish data are shown in
Table 2. To Dixons classification, an additional type, termed social categories
was added, in part following Silva-Corvaln (1994), to account for adjectives that
refer to nationality (cubano Cuban), race (blanco white), social status (rico
rich), and other adjectives used to socially categorize individuals. In SilvaCorvalns classification racial color terms were not considered separately from

10

Gabriela G. Alfaraz

color adjectives; here, however, the category color only included inanimate objects,
and racial terms that refer to skin color are classified as social categories, similar to
nationality and socioeconomic status.
Dimension
Age
Value
Color
Physical
property
Human
propensity
Speed

big, long, tall, wide, deep

new, young, old
good, lovely, atrocious,
perfect, strange, important
black, white, red
hard, heavy, clean, hot, sour

grande big, alto tall

nuevo new, viejo old
bueno good, bello beautiful,
cmico funny
negro black, morado purple
suave soft, slido solid

jealous, happy, kind,

clever, generous, proud
fast, quick, slow

inteligente intelligent, amable

kind, travieso mischievous
rpido fast, brusco sudden

Table 2. Adjective types (Dixon 2004, pp.3-4)

The analysis included 601 tokens of the copulas with predicate adjectives.
Tokens were counted if the context clearly indicated a class rather than individual
reference; thus, tokens such as estar in (4), in which the speaker used it to describe
her current physical appearance compared to a previous one, were not included
because the comparison is with the woman herself at different points in time. The
data analysis and coding involved two stages of reliability calculations. As a first
step, a determination was made that estar occurred in contexts in which ser could
occur and that there was not a comparison of the entity to itself; an intercoder
reliability score of 98.7% was obtained for two coders. In the second step, the
adjectives were coded in the eight categories. An initial intercoder reliability score
of 97.3% was reached and disagreements were resolved to arrive at 100%
agreement.
(4) Y yo le dije no yo estoy muy fea. ... Digo no, no tengo muchas canas y estoy
muy fea.
I told him no I am too ugly... I said no, no I have too much gray hair and I am
too ugly.
A regression analysis was carried out with Goldvarb, a program designed for
the analysis of linguistic variation. The variables included in the analysis were
copula, generation/time, gender, and type of adjective. Although socioeconomic
status typically renders valuable information about the status of linguistic variation
in a community, it was not considered because the speakers were middle class and
there was not enough information about their work and educational backgrounds to

The Status of the Extension of estar in Cuban Spanish

11

establish differences in their position within the middle class, i.e. lower, middle or
upper. The dependent variable was the type of copula. Goldvarb generates results in
the form of probability weights that range from 1 to 0, which indicate the strength of
each factor (i.e. male vs. female) in a factor group (i.e. gender). The closer the
weight is to 1, the greater the strength of the factor, and the closer to 0, the weaker
its strength. Weights clustering around .5 indicate a neutral contribution to the
variation. Furthermore, a weight of 1 indicates categorical occurrence and 0
categorical absence of the variant, in other words, there is no variation. In addition
to the probability weights, a range score is reported that indicates the strength of
each factor group on the variation being studied. For the results presented here, the
probability weights are for the presence of estar.
3.

Results

Variation of the copulas preceding predicate adjectives was found for the three
groups included in the study. Innovative estar occurred in 19.3% (116) of the tokens
(601) examined for the three groups. As evident from the results in Table 3, in
which the probability weight is for the extension of estar, the factors that were
significant in the multivariate analysis were generation/time and adjective type;
gender, however, was not significant. The strength of the two significant factors, as
indicated by their range scores, 27 generation/time and 29 adjective type, is similar,
which suggests that both age and adjective type contribute more or less equally to
the variation. These findings differ from those reported in Corts-Torres (2004) and
Salazar (2007), in which adjective type had the strongest influence on the variation
and age was not significant in conditioning the variation. Each of the factor groups
shown in this table will be discussed in turn below.
Group
Factor
Generation/ 1960s
time
1990s Older
1990s Younger
Range
Gender*

Weight
.38

%
13

N
34/262

.46

14.8

17/115

.65

29

65/224

27

Female

.49

17.5

65/371

Male

.52

22.2

51/230

Range

12

Adjective
Type

Gabriela G. Alfaraz

Dimension

.46

20

14/70

Age

.39

18.2

4/22

Value

.56

26.4

71/269

Color

0/23

Physical property

.59

26.6

17/64

Human propensity

.3

11.3

9/80

Speed

0/7

Social category

0/65

Range
29
*Factor group not significant. N = 601; Input .21; Chi-square per cell
.90; Log likelihood -257.083
Table 3. Results (estar is the application value)
3.1 Gender
Gender commonly plays a role in linguistic changes, according to Labov (2001),
who pointed out that gender is a differentiating factor in almost every case of
stable social stratification and change in progress that has been studied (p.262). As
a general trend, females, although they tend to be more conservative than their male
counterparts, often lead linguistic changes, whether from below or from above. For
the extension of estar, however, gender was not a significant factor, as shown in
Figure 1. The probability weights for males and females were essentially equal, .52
and .49, and neutral, as they cluster around .5. Frequency rates showed a small
difference between the genders the male group had a slightly higher percentage of
estar than females (22.2% vs. 17.5%), but again, this difference was not significant.
These results are consistent with other studies that examined the degree of
influence of gender on the extension of estar. In her study in Cuernavaca, Mexico,
Corts-Torres (2004) also found that gender was not a significant factor. The rate of
innovative estar was 50% for both males and females. In New Mexican Spanish,
Salazar (2007) similarly found that gender was not significant although rates of
innovative estar were higher for males (54%) than females (46%), which would be
expected for a nonstandard variant. Thus, even though gender commonly influences
variation, the extension of estar is not conditioned by it.

The Status of the Extension of estar in Cuban Spanish

13

1
0.9
0.8
Probability

0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
Female

Male

Figure 1.
Results for estar by gender. Factor group was not significant.
Probability weights: Female .49, male .52. Percentages: Female 17.5%
(65/371); male 22.2% (51/230). (N = 601; Input .21; Chi-square per cell .90;
Log likelihood -257.083)
3.2 Generational groups in real and apparent time
3.2.1

Same generation in real time

The results for the generational cohort recorded thirty years apart, in the late 1960s
and 1990s, revealed that there is extension of estar at both time periods, and that in
spite of the thirty-year span in the time the data were collected, the behavior of the
two groups was relatively similar, as indicated by a small difference in the
probabilities, .38 for the 1960s and .46 for the 1990s. Both groups disfavored estar,
as shown in Figure 2, although the probability for the 1990s group showed a smaller
effect. The differences, in terms of frequency, are small, less than 2% between the
two groups (13% and 14.8%). Sankoff & Wagner (2006) noted in their real time
study of the inflected future in Montral French that when there is community
change, older speakers may show differences in real time that result from their
response to the higher frequency in younger speakers of a variant involved in
change; in Montral, the frequency of the new variant was related to socioeconomic
status (SES), with speakers with the highest SES reducing its frequency as they
aged and those with the lowest SES showing a flat trajectory over time. The
difference in the use of estar in the two time periods could be explained as a

14

Gabriela G. Alfaraz

consequence of older speakers following younger ones, but the fact that the
difference between the two time periods is very small, suggests instead that the
extension of estar remained stable for these speakers as they advanced to old age.
1
0.9
0.8
Probability

0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
1960s

1990s Older

Figure 2.
Same generation in real time. Probability of estar for same
generational cohort in 1960s and 1990s. Probability weights: 1960s .38, 1990s
Older .46. Percentages: 1960s 13% (34/262) and 1990s 14.8% (17/115). (N =
601; Input .21; Chi-square per cell .90; Log likelihood -257.083)
Furthermore, the small difference found in the comparison of this generational
cohort as younger and older adults is evidence that the extension of estar does not
reflect age-grading, a pattern of variation in which individuals alter their linguistic
behavior as they progress through the lifespan (Labov 1994). Age-grading is
commonly found with stable variants whose frequency reflects different life stages
(e.g. youth, adulthood, old age). Evidence of a change in progress is based on an age
cohort maintaining the same frequency as they advance from adolescence to young
adulthood, middle and old age. In generational change, it is not the individuals in
the cohort that change, but rather the community (Labov 2001). When there is
change, whether from above or below, younger speakers tend to have a different
frequency than the older ones. The real-time results found for the older generational
cohort set the ground for interpreting a change in the frequency of estar as
generational change because the linguistic behavior of the older group was shown to
vary only minimally over time, but findings for the younger generation of Cubans
are needed to add to the picture; if a higher frequency of the form is found for the
1990s younger group, then it adds evidence to suggest a change in progress.

The Status of the Extension of estar in Cuban Spanish

3.2.2

15

Different generation in real time

The real-time analysis of the speakers recorded at the same age at different points in
time, 1960s and 1990s, who represent different generations, showed significant
differences in their use of estar. These results are shown in Figure 3. The
probability weight for the 1990s younger group was .65, which suggests a tendency
toward the extension of estar. The older generation, on the other hand, had a
probability weight of .38, which shows that it disfavored estar. The frequency of
estar for the younger generation was slightly more than double that of the older one,
29% for the 1990s younger and 13% for the 1960s group.
The different tendencies observed for the two generational groups in real time
suggest a change in progress. The younger generation outdoes the older one in the
frequency of the extension of estar. The advancement of the change appears to have
been rapid, as indicated by the fact that the percentage rates of estar had doubled in
the younger group. While it cannot be concluded that this is a change in progress
without studying even younger speakers, the evidence found here for these two
generations points to a change from below in the speech community that parallels
the changes reported for other varieties (Gutirrez 1992, 1994, Silva-Corvaln 1986,
1994). What is made clear with these findings is that the extension of estar is not
stable in this variety but has been advancing in younger generations. The study of
speakers younger than these young adults in the 1990s would add valuable
information about the distribution across age groups of the change.
1
0.9
0.8
Probability

0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
1960s

1990s Younger

Figure 3.
Different generations in real time. Probability of estar for 1960s
and 1990s Younger. Probability weights: .38 1960s and .65 1990s. Percentages:

16

Gabriela G. Alfaraz

1960s 13% (34/262) and 1990s Younger 29% (65/224). (N = 601; Input .21; Chisquare per cell .90; Log likelihood -257.083)
3.2.2

Different generation in apparent time

The results for the analysis of different generations in real time pointed to
generational change; it was also shown that the cohort of speakers studied at two
points in time did not change significantly over the span of time studied here. The
comparison of the older and younger generations in apparent time, then, is expected
to corroborate the findings in real time. The results for the 1990s Older and
Younger groups in Figure 4 show probability differences similar to those for the
generational comparison in real time. The 1990s older group had a probability
weight of .46 and the younger group one of .65. In terms of the frequency of
innovative estar, for the younger group it was nearly double the rate found for the
older group (14.8% older and 29% younger).
1
0.9
0.8
Probability

0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
1990s Older

1990s Younger

Figure 4.
Results for different generations in apparent time. Probability of
estar for 1990s Older and Younger. Probability weights: .46 Older and .65
Younger. Percentages: Older 14.8% (17/115); Younger 29% (65/224). (N = 601;
Input .21; Chi-square per cell .90; Log likelihood -257.083)
Studies that similarly examined the extension of estar in apparent time,
including Corts-Torres (1994) and Salazar (2007), did not find that age was a
significant factor contributing to the occurrence of estar. Corts-Torres found that
the middle age group had higher rates of estar than the older or the younger; thus,

The Status of the Extension of estar in Cuban Spanish

17

the extension of estar appears to be a stable, age-graded feature in the variety of

Mexican Spanish she studied. This is in contrast to these findings for Cuban
Spanish, in which generation was significant and the probability weights for
generational groups suggested that the younger generation favored and the older
disfavored the extension of estar.
3.3 Adjective type
The analysis of internal constraints on the variation showed that the extension of
estar is not progressing at the same rate through the various adjective contexts
studied. The results of the Goldvarb analysis in Figure 5 show that there is no
evidence that estar has extended to adjectives of color, social category or speed, as
these types categorically occurred with ser. At the same time, it was found that
adjectives of human propensity disfavored estar, with a probability weight of .30.
Human propensity adjectives include amable kind, carioso affectionate,
fantico fanatic, and romntico romantic. The frequency of estar with these
adjectives was 11%. Another adjective type that disfavored the extension of estar,
with a probability weight of .39, included age adjectives, including joven young,
antiguo old, and viejo old. The rate of estar for these age adjectives was 18.2%.
The last adjective type that disfavored estar, although only slightly (.46), included
dimension adjectives, such as alto tall, chiquito small, corto short, and estrecho
narrow; the frequency of these adjectives was 20%. There were two adjective
types, physical property and value, in which the extension of estar was favored, but
the probability weights indicate a small influence. Physical property adjectives,
including firme firm, fresco fresh, and dbil weak, showed a small preference
for estar, probability weight of .59 and rate of 26.6%. Similarly, value adjectives,
including asombroso amazing, gracioso funny/pretty, lindo pretty, interesante
interesting, horrible horrible, chulo cute, and fcil easy were found to slightly
favor estar, probability weight of .56 and rate of 26.4%.

Gabriela G. Alfaraz

ry
ca
te
go

Sp
ee
d

So
ci
al

ity
pr

an
H
um

sic
al

pr

op

op

en
s

er
ty

lo
r
Co
Ph
y

V
al
ue

D
im
en

ge

1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
sio
n

Probability

18

Figure 5.
The probability of estar by adjective class. Probability weights:
Dimension .46, Age .39, Value .56, Color .0, Physical property .59, Human
propensity .30, Speed .0, Social category .0. Percentages: Dimension 20%
(14/70), Age 18.2% (4/22), Value 26.4% (71/269), Color 0% (0/23), Physical
property 26.6% (17/64), Human propensity 11.3% (9/80), Speed 0% (0/7),
Social category 0% (0/65). (N = 601; Input .21; Chi-square per cell .90; Log
likelihood -257.083)
These results confirm that the extension of estar has not advanced to all
adjective classes; in fact, six of the eight adjective types examined did not favor the
variant estar, either because there was categorical nonoccurrence or because the
probability weight indicated a tendency to disfavor it. Adjective types in which no
cases of the extension of estar were found included speed, color, and social
categories, as shown in (5). In (a) there is a speed adjective with ser. Estar did not
occur with color adjectives, as evident in (b), a description of the colors of autumn
leaves encountered during one subjects trip to Toronto and Ottawa, Canada, in
which the changing quality of the colors did not trigger the use of estar, es rojo its
red and es amarillo its yellow. An evaluative remark about the experience,
however, occurred with estar, est divino its divine. Lastly, adjectives that
referred to social categories were categorical nonoccurrences; an example is shown
in (c).
(5) Categorical nonoccurrence
a. Ella es rpida en la mquina.
She is fast on the machine. (1960s)

The Status of the Extension of estar in Cuban Spanish

b.

c.

19

Ay los rboles son en esa poca...nosotros fuimos en octubre y los rboleshay veces que usted vea un rbol al lado del otro y no tienen- dos hojas no
tienen el mismo color....un rbol es rojo y el otro que tiene al lado es
amarillito y el otro verde. Est divino. Divino, divino.
Oh the tree during that season...we went in October and the treessometimes you see one tree next to another and they dont have- two
leaves arent the same color...one tree is red and the other that is next to it
is yellow and the other green. It is divine. Divine, divine. (1960s)
Yo era comunista porque lo tena que ser obligatoriamente.
I was communist because I was obliged to be. (1990s Younger)

The adjective contexts that disfavored the extension of estar included those in
which the adjective denoted human propensity, age, or dimension. Examples of the
use of estar are shown in (6). Estar was strongly disfavored with adjectives
denoting human propensity, such as carioso affectionate, caritativo charitable,
independiente independent, and agradable kind; (a) shows an example of estar
with the adjective despabilado smart/quick in which the speaker expresses her
ideas about things that appeal to smart or bright children. Extension of estar was
also disfavored with age adjectives; an example of estar with an age adjective is
shown in (b), which refers to men who are old as a collective. Dimension adjectives,
such as grande big and largo long, were also found to disfavor estar; one
example of its use is shown in (c), where estar is used by a 1960s subject to
describe the length of a miniskirt worn by her boss wife.
(6) Disfavor extension of estar
a. Esos nios que estn despabilados, les gustan porque los que no les gustan
no.
Those kids that are smart/quick, they like them because those that dont
like them no. (1990s Younger)
b. Hay hombres que estn viejos, que usted los oye hablar y se le saltan las
lgrimas.
There are men that are old, that you hear them talk and tears come to your
eyes. (1990s Younger)
c. [La falda] estaba demasiado corta... es muy exagerada la moda...yo creo
que no es elegante la moda tan exagerada.
[The skirt] was too short...the fashion is too extreme...I dont think fashion
that extreme is elegant. (1960s)
The two adjective types value and physical property showed a small tendency
to favor estar; examples are shown in (7). Adjectives of physical property appeared
with estar, as in (a) in which fuerte strong refers to the strength or sturdiness of a
boat on which the male speaker from the 1990s younger group fled Cuba. Value

20

Gabriela G. Alfaraz

adjectives similarly occurred with estar, as in (b) in which estar was used to
describe paintings by a Cuban artist that the speakers saw at a fair in Miami in the
1960s, or in (c), from the 1990s older group, in which estar was used to describe a
woman in a picture whom the speaker had never seen before. Other examples of
value adjectives with estar include the parallel examples in (d) from the 1960s data
and (e) from the 1990s younger generation.
(7) Favor extension of estar
a. El barco estaba fuerte.
The boat was strong/sturdy. (1990s Younger)
b. Estaban bonitos los cuadros que hizo la seora de ( )...los de los pjaros
estaban lindos, me gustaron.
The painting that the woman from ( ) made were pretty... the ones with
the birds were pretty, I liked them. (1960s)
c. [La mujer] no est fea. Tiene facciones bonitas.
[The woman] is not ugly. She has pretty features. (1990s Older)
d. La pelcula est simptica. (1960s)
The movie is amusing.
e. La pelcula est cmica.
The movie is funny. (1990s Younger)
These findings for adjective type when compared to other studies reveal
similarities and differences in the rates of estar according to this factor group.
Adjectives of age were found to favor the extension of estar in Corts-Torres (2004)
and Salazar (2007), but here they disfavored it. The rate of estar with age adjectives
reported in Gutirrez (1992, 2003) was considerably higher, compare 43% to 18%,
but Ortiz Lpezs findings for San Juan indicated a lower rate (15%). Higher rates
for age adjectives were also reported in de Jonge (1993) for Mexico City (35%) and
Caracas (53%) and Silva-Corvaln (1994) reported significantly higher rates (78%).
Differences are also evident in the findings for dimension adjectives, which were
slightly disfavored here as in Corts-Torres, but the 20% rate found here is lower
than the rates of 34% reported in Gutirrez (1992) and 36% reported in Ortiz Lpez
(2000). Adjectives denoting properties classified as human propensity have low
rates of estar across varieties. Corts-Torres (2004) showed that similar adjectives
disfavor estar more strongly than in Cuban Spanish, as indicated by the probability
weight of .19. In general, studies have found low rates of estar with adjectives of
this type; Gutirrez (1992) reported 2% estar for adjectives such as honesto honest
and respetable respectable, Ortiz Lpez found 3.4%, and Silva-Corvaln (1994)
reported only 13%, a low rate considering the high frequency of estar she found
overall. In the varieties in which there is extension of estar, whether it has been
described as stable variation or a change in progress, the internal linguistic factor,
type of adjective, is an important factor constraining the variation.

The Status of the Extension of estar in Cuban Spanish

4.

21

Discussion

The quantitative findings presented above both confirmed the extension of estar
preceding predicate adjectives in Cuban Spanish and suggested a change in progress
in this variety. The overall frequency found here of 19.3% for conversational data is
higher than the rates reported for data from conversations in other studies. Gutirrez
(1994), for instance, reported a rate of 16% for conversations and 24% for
questionnaires, Ortiz Lpez (2000) reported rates of 12% for conversations and 19%
for questionnaires for Puerto Rican Spanish, Corts-Torres (2004) reported a
frequency of 23% overall, but only 9% for conversations, and 31% for
questionnaires; Silva-Corvaln (1994) found a higher rate of 34% for conversational
data, following the general trend to find high rates of estar for the speakers in her
sample.
The comparison of generational groups in real and apparent time revealed a
significant change in the frequency of estar, with an increase in its frequency in the
younger generation. The logistic regression analysis showed that the younger group
had a tendency to favor estar, whereas the generational cohort recorded in the 1960s
and 1990s tended to disfavor it. These findings were substantiated in both the
studies in real and apparent time. A comparison of the two groups that represent the
older generation revealed a similar pattern for the extension of estar, in spite of the
near thirty-year span in the time the data was collected, suggesting that the variation
of the copulas had remained relatively stable for this generation. As noted in
Sankoff & Wagner (2006), linguists operate under the assumption that grammar
does not change after the critical period of acquisition (Lenneberg 1967). These
findings for estar provide evidence to support the idea that after the frequency of
variants according to the social and linguistic constraints that determine the
variation in the speech community at the specific time is set it will remain stable
over the course of the lifespan.
Findings from the real and apparent time studies suggested that estar has been
advancing in Cuban Spanish. Gutirrez (1992) proposed that in the Spanish of the
community he studied in Morelia, Michoacn, Mxico, estar was a change in
progress. Other studies, however, have not found evidence of change. As noted
above, Corts-Torres (2004) did not find that age was a significant factor in
Cuernavaca, Mexico; de Jonge (1993) concluded, after comparing results for age
adjectives from the late 1960s and late 1980s from Mexico City, that the
progression of estar appeared to be stable in Mexican Spanish; and de Jonges realtime comparison of the use of estar with age adjectives in data from Caracas,
Venezuela from the early 1970s and 1980s also failed to indicate a change in
progress, at least with that type of adjective, but perhaps looking at a wider range of
adjectives would have uncovered some evidence of change. As de Jonge noted, lack
of evidence of change over time does not indicate that a change is not taking place.

22

Gabriela G. Alfaraz

Indeed, linguistic changes may remain stable for hundreds of years without
advancing to completion (Labov 2001).
Additional studies on the extension of estar are necessary to estimate its degree
of advancement with different adjective types. The explanatory power of
variationist studies can be expanded with the inclusion of findings from
grammatical acceptability judgments (Henry 1995, 2005) and other tasks that may
render information about the social and linguistic constraints on the variation
(Alfaraz 2010). For Cuban Spanish, further study with a sample of younger speakers
and a range of socioeconomic status groups is necessary to determine whether estar
is indeed a change in progress and whether it continues to advance in this variety.
5.

Conclusion

Although the occurrence of ser and estar with predicate adjectives has been
extensively discussed in Spanish linguistics, there are few variationist studies that
have centered on the social and linguistic distribution of the copulas to examine the
claim that the copulas have been undergoing change in which estar has been
generalizing to contexts previously reserved for ser, resulting in variation between
the forms in adjective contexts. The research reported on in this paper examined the
variation of ser and estar in Cuban Spanish in order to determine whether there is
extension of estar in this variety, to describe the social and linguistic factors that
influence the variation, and to explore whether its distribution suggested stable
variation or a change in progress.
These questions were studied using both a real-time and an apparent-time
analysis. The real-time study was carried out with data from the 1960s and the
1990s and it involved comparing a generational cohort at the two points in time to
compare its use of estar over time and it also involved comparing this generational
cohort with a younger one that was recorded in the 1990s. The apparent time study
compared two generational groups, an older and a younger one, at the same time
period. The results indicated that the frequency of estar in the generational cohort
compared at two points in time had remained stable and that the younger generation
showed a tendency to favor estar. These findings confirmed that there is extension
of estar in this variety and they pointed to a change in progress. The ranking of
constraints showed that the variation was influenced more or less equally by
generation and adjective type but that gender was not a significant factor.
The findings for adjective type indicated that the diffusion of estar was
influenced by the linguistic context and that only two of the eight adjectives types
studied favored the extension of estar; three adjective types had probability weights
indicating that they disfavored estar, and three other adjective types were
categorical nonoccurrences.
The question of whether the copula estar preceding adjectives has continued to
advance is one that has not received attention in many varieties of Spanish,

The Status of the Extension of estar in Cuban Spanish

23

including the variety reported on here. The study of the copulas using a variationist
approach, however, can render accounts of the variation of the copulas that can be
viewed together to understand stages of the change in which estar appears to be
involved.
References
Alfaraz, Gabriela G. 2010. The influence of social factors on Spanish dialect contact
in the US: A look at Mexican and Cuban Spanish in a Midwestern city.
Southwest Journal of Linguistics 29(2), 27-34.
Alonso, Amado & Pedro Henrquez Urea. 1939. Gramtica castellana. Buenos
Aires: Losada.
Andrade, Manuel J. 1919. The distinction between ser and estar. Hispania 2, 19-23.
Bailey, Guy, Natalie Maynor & Patricia Cukor-Avila (eds.). 1991. The emergence
of Black English: Texts and commentary. Amsterdam: John Benjamins.
Bello, Andrs. 1847. Gramtica: gramtica de la lengua castellana destinada al uso
de
los
americanos.
Available
at
http://www.cervantesvirtual.com/servlet/SirveObras/121450742290360515434
35/index.htm.
Bolinger, Dwight L. 1947. Still more on ser and estar. Hispania 30, 361-367.
Buchstaller, Isabelle. 2009. The quantitative analysis of morphosyntactic variation:
Constructing and quantifying the denominator. Language and Linguistic
Compass 3/4, 1010-1033.
Bull, William E. 1942. New principles for some Spanish equivalents of to be.
Hispania 25, 433-443.
Cedergren, Henrietta. 1973. The interplay of social and linguistic factors in Panama.
Doctoral dissertation, Cornell University.
Clements, Joseph Clancy. 1988. The semantics and pragmatics of the Spanish
COPULA + ADJECTIVE construction. Linguistics 26, 779-822.
Corts-Torres, Mayra. 2004. Ser o estar? La variacin lingstica y social de estar
ms adjetivo en el espaol de Cuernavaca, Mxico. Hispania 85, 788-795.
Crespo, Luis. 1946. Los verbos ser y estar explicados por un nativo. Hispania 29,
45-55.
Dixon, Robert M.W. 2004. Adjective classes in typological perspective. In Robert
M.W. Dixon & Alexandra Y. Aikhenvald (eds.), Adjective classes: A crosslinguistic typology, 1-49. Oxford: Oxford University Press.
Falk, Johan. 1979. SER Y ESTAR con atributos adjetivales. Anotaciones sobre el
empleo de la cpula en cataln y en castellano, I. Studia Romanica Upsaliensia
29. Stockholm: Almqvist & Wiksell.

24

Gabriela G. Alfaraz

Franco, Fabiola & Donald Steinmetz. 1983. Ser y estar + adjetivo calificativo en
espaol. Hispania 66, 176-184.
Franco, Fabiola & Donald Steinmetz. 1986. Taming ser and estar with predicate
adjectives. Hispania 69, 377-386.
Gutirrez, Manuel J. 1992. The extension of estar: A linguistic change in progress in
the Spanish of Morelia, Mxico. Hispanic Linguistics 5, 109-141.
Gutirrez, Manuel J. 1994. Ser y estar en el habla de Michoacn, Mxico. Mxico,
DF: Universidad Autnoma de Mxico.
Gutirrez, Manuel J. 2003. Simplification and innovation in US Spanish.
Multilingua 22, 169-184.
Henry, Alison. 1995. Belfast English and standard English: Dialect variation and
parameter setting. Oxford: Oxford University Press.
Henry, Alison. 2005. Non-standard dialects and linguistic data. Lingua 115, 15991617.
Jonge, Bob de. 1993. (Dis)continuity in language change: ser and estar + adjective
in Latin American Spanish. Linguistics in the Netherlands 10, 69-80.
Labov, William. 1963. The social motivation of a sound change. Word 19, 273-309.
Labov, William. 1966. The social stratification of English in New York City.
Washington, DC: Center for Applied Linguistics.
Labov, William. 1972. Sociolinguistic patterns. Philadelphia: University of
Pennsylvania Press.
Labov, William. 1978. Where does the linguistic variable stop? A response to
Beatriz Lavandera. Working Papers in Sociolinguistics 44. Austin: Southwest
Educational Development Laboratory.
Labov, William. 1994. Principles of linguistic change. Vol. 1: Internal factors.
Malden, MA: Blackwell.
Labov, William. 2001. Principles of linguistic change. Vol. 2: Social factors.
Malden, MA: Blackwell.
Lenneberg, Eric Heinz. 1967. Biological Foundations of Language. New York:
Wiley.
Lujn, Marta. 1981. The Spanish copulas as aspectual indicators. Lingua 54, 165210.
Mello, George de. 1979. The semantic values of ser and estar. Hispania 62, 338341.
Navas-Ruiz, Ricardo. 1963. Ser y estar: estudio sobre el sistema atributivo del
espaol. Salamanca: Universidad de Salamanca.
Ortiz Lpez, Luis A. 2000. Extensin de estar en contextos de ser en el espaol de
Puerto Rico: evaluacin interna y/o contacto de lengua. BAPLE, 99-118.
Pountain, Christopher. 1982. Essere/stare as a Romance phenomenon. In Nigel
Vincent & Martin Harris (eds.), Studies in the Romance verb, 139-160.
London: Croom-Helm.

The Status of the Extension of estar in Cuban Spanish

25

Roldn, Mercedes. 1974. Toward a semantic characterization of ser and estar.

Hispania 57, 68-75.
Salazar, Michelle L. 2007. Est muy diferente a como era antes: Ser and Estar +
adjective in New Mexico Spanish. In Kim Potowski & Richard Cameron (eds.),
Spanish in contact: Policy, social and linguistic inquiries, 343-353. Amsterdam,
Philadelphia: John Benjamins.
Sankoff, Gillian & Suzanne Evans-Wagner. 2006. Age-grading in retrograde
movement: The inflected future in Montral French. U Penn Working Papers in
Linguistics 12, 1-14.
Sera, Maria D. 1992. To be or to be: Use and acquisition of the Spanish copulas.
Journal of Memory and Language 31, 408-427.
Silva-Corvaln, Carmen. 1986. Bilingualism and language change: The extension of
estar in Los Angeles Spanish. Language 62, 587-608.
Silva-Corvaln, Carmen. 1994. Language contact and change: Spanish in Los
Angeles. Oxford & New York: Oxford University Press.
Va-Cerd, Antonio. 1982. Ser y estar + adjetivos: un estudio sincrnico y
diacrnico. Tbingen: Gunter Narr Verlag.

Copyright of Studies in Hispanic & Lusophone Linguistics is the property of Studies in Hispanic & Lusophone
Linguistics and its content may not be copied or emailed to multiple sites or posted to a listserv without the
copyright holder's express written permission. However, users may print, download, or email articles for
individual use.