ArticlePDF Available

Can quirks of grammar affect the way you think? Grammatical gender and object concepts



Can the language you speak affect the way you think? Unlike English, many languages have a grammatical gender system whereby all nouns are assigned a gender. Does talking about inanimate objects as if they were masculine or feminine actually lead people to think of inanimate objects as having a gender? A series of studies found effects of grammatical gender on people's perceptions of similarity between objects and people. This was true even though the tasks were performed in English (a language devoid of grammatical gender), even when the tasks were non-linguistic (e.g., rating similarities between unlabeled pictures), and even while subjects were engaged in a verbal interference task. Finally, results showed that cross-linguistic differences in thought can be produced just by grammatical differences and in the absence of other cultural factors.
Can Quirks of Grammar Affect the Way You Think?
Grammatical Gender and Object Concepts
Webb Phillips (
MIT NE20-457, 77 Mass Ave
Cambridge, MA 02139 USA
Lera Boroditsky (
MIT NE20-456, 77 Mass Ave
Cambridge, MA 02139 USA
Can the language you speak affect the way you think? Unlike
English, many languages have a grammatical gender system
whereby all nouns are assigned a gender. Does talking about
inanimate objects as if they were masculine or feminine
actually lead people to think of inanimate objects as having a
gender? A series of studies found effects of grammatical
gender on people’s perceptions of similarity between objects
and people. This was true even though the tasks were
performed in English (a language devoid of grammatical
gender), even when the tasks were non-linguistic (e.g., rating
similarities between unlabeled pictures), and even while
subjects were engaged in a verbal interference task. Finally,
results showed that cross-linguistic differences in thought can
be produced just by grammatical differences and in the
absence of other cultural factors.
Humans communicate with one another using a dazzling
array of languages, and each language differs from the next
in innumerable ways (from obvious differences in
pronunciation and vocabulary to more subtle differences in
grammar). For example, to say that “someone ate the
cheese” in English, we must include tense - the fact that the
event happened in the past. In Russian, the verb would also
have to include whether the cheese-eater was male or
female, and whether said cheese-eater ate all of the cheese
or just a portion of it. Speakers of different languages have
to attend to and encode strikingly different aspects of the
world in order to use their language properly (Sapir, 1921;
Slobin, 1996).
Yet despite all these differences, speakers of all languages
inhabit very similar bodies, and need to communicate about
very similar physical worlds. This has lead some scholars to
argue that speakers of different languages vary only in their
speaking, and not in their thinking. Beyond the surface
structure of languages, the argument goes, lies a universal
language of thought (Fodor, 1975).
The question of universality of mental representations
(whether or not speakers of different languages think
differently about the world) has long been at the center of
controversy attracting scholars from Plato to Chomsky, but
despite much attention and debate, definitive answers have
not been forthcoming. Findings presented in this paper
suggest that people’s mental representations of the world are
not universal. Even people’s ideas about concrete objects
can be shaped by implementational quirks of their
Grammatical Gender
Unlike English, many languages have a grammatical gender
system whereby all nouns (e.g., penguins, pockets, and
toasters) are assigned a gender. Many languages only have
masculine and feminine genders, but some also assign
neuter, vegetative, and other more obscure genders. When
speaking a language with grammatical gender, speakers are
required to mark objects as gendered through definite
articles, gendered pronouns, and often need to modify
adjectives or even verbs to agree in gender with the nouns.
Could the grammatical genders assigned to objects by a
language influence people’s mental representations of
Forks and frying pans do not (by virtue of being
inanimate) have a biological gender. The perceptual
information available for most objects does not provide
much evidence as to their gender, and so conclusive
information about the genders of objects is only available in
language (and only in those languages that have
grammatical gender). It is possible that language has the
greatest influence on thought in abstract domains like
grammatical gender ones not so reliant on sensory
experience (Boroditsky, 2000; 2001). For example, people’s
subjective conception of time (say as a vertical or a
horizontal medium) is not constrained by sensory
experience, and appears to vary across languages and
cultures (see Boroditsky, 2000; 2001 for further discussion).
This paper examines whether people’s mental
representations of objects are influenced by the grammatical
genders assigned to the objects names in their native
But Isn’t Grammatical Gender Arbitrary?
The assumption that grammatical gender has no meaning is
widespread in the field. For example, the purely
grammatical nature of grammatical gender has been a key
(though untested) assumption in the recent debate about the
tip-of-the-tongue phenomenon (Caramazza & Miozzo,
1997; Vigliocco et al., 1997). Indeed, a priori, there are
several reasons to think that people would not take
grammatical gender as meaningful. First, the assignment of
grammatical gender to object names often appears to be
semantically arbitrary (and sometimes downright absurd).
As Mark Twain noted, “In German, a young lady has no
sex, while a turnip has…a tree is male, its buds are female,
its leaves are neuter; horses are sexless, dogs are male, cats
are female... tomcats included.” Further, the grammatical
genders assigned to names of particular objects vary greatly
across languages (Braine, 1987). For example, the sun is
feminine in German, but masculine in Spanish, and neuter in
Russian. The moon, on the other hand, is feminine in
Spanish and Russian, but masculine in German.
On the other hand, there are also reasons why we might
expect people to take grammatical gender as meaningful.
First, since many other grammatical distinctions reflect
differences that are observable in the world (the plural
inflection, for example), children learning to speak a
language with a grammatical gender system have no a priori
reason to believe that grammatical gender doesn’t indicate a
meaningful distinction between types of objects. Indeed,
many adult philosophers throughout history have thought
that grammatical gender systems were a reflection of the
essential properties of objects, and even took a considerable
amount of pride in the thought that the natural genders of
objects would be captured in the grammatical subtlety of
their language (Fodor, 1959). Children learning a language
may make similar (though perhaps less patriotically minded)
Second, since most children grow up learning only one
language, they have no opportunity to perform the
comparative linguistics necessary to discover the seemingly
arbitrary nature of grammatical gender assignment. For all
they know, the grammatical genders assigned by their
language are the true universal genders of objects. Finally,
speakers of languages with grammatical gender must mark
gender almost every time they utter a noun (hundreds or
thousands of times a day). The sheer weight of repetition (of
needing to refer to objects as masculine or feminine) may
leave its semantic traces, making the objects’ masculine or
feminine qualities more salient in the representation.
Can Grammatical Gender Affect Concepts?
We hypothesize that the grammatical genders assigned to
objects by a language influence people’s mental
representations of objects.
Experiment 1: Picture Similarity
Experiment 1 was designed to test whether an objects’ name
being grammatically feminine or masculine in a language
leads speakers of that language to think of the object itself
as more like a male or female. Spanish and German
speakers were asked to rate the similarity of objects and
animals to human males and females. All subjects were
tested in English, all items were presented as unlabeled
pictures, and all of the objects and animals chosen for this
study had opposite grammatical genders in Spanish and
German. We hypothesized that same-gender pairs would be
rated as more similar.
Participants Twenty-two Spanish-English bilinguals, and
thirty-three German-English bilinguals participated in the
study on a volunteer basis. Participants ranged in age from
17 to 69 years (M=32.9 yrs old). All of the Spanish and
German speakers considered Spanish and German
(respectively) to be their native language, but both groups
were highly proficient in English and had an average of 15.9
years of experience with English (16.4 years for German
speakers, and 15.4 years for Spanish speakers).
Materials Materials used in this study comprised 14
pictures of objects and animals, and 8 pictures of people. Of
the pictures of people, 4 were pictures of females (a woman,
a ballerina, a bride, and a girl), and 4 were pictures of males
(a man, a king, a giant, and a boy). Of the objects, half were
items whose names are masculine in German but feminine
in Spanish (toaster, moon, spoon, broom, whale, frog, fox),
and half were items whose names are feminine in German,
but masculine in Spanish (clock, sun, fork, toothbrush,
mouse, snail, cat). The pictures of objects and animals were
chosen such that they had a single dominant label in both
Spanish & German which had opposite genders in the two
Each participant provided a similarity rating for every
possible person—object comparison (a total of 112) on a
scale of 1 (not similar) to 9 (very similar). A new random
order for the comparisons was generated for each subject.
Spanish and German speakers completed the same
experimental task, and both groups completed the task in
English. Participants read the following instructions “In this
study, you will see pairs of pictures appear on the screen. In
each pair, there will be a picture of a person on the left and a
picture of an object or animal on the right. Your task is to
tell us how similar you think the two things being depicted
are. You will see a scale where 1=not similar and 9= very
similar. For each pair of pictures, please choose a number
between 1 and 9 to indicate how similar you think the two
things are. Please use the whole scale (give some 1's and
some 9's and some of all the numbers in-between).”
Procedure Participants were tested individually. A
computer presented the experimental materials and recorded
the participants’ responses. Each pair of pictures remained
on the screen until a participant made a response.
Participants made responses by clicking on one of the boxes
(numbered 1 through 9) on the screen to signify a similarity
Results and Discussion Subjects found greater similarity
between people and objects of matching gender than
between people and objects of non-matching gender (by
subjects: t = 3.08, df = 42, p < .0018; by items: t = 4.62, df =
7, p = .0012). This suggests that Spanish and German
speakers indeed end up thinking about objects as more
similar to biological males and females, depending on the
object’s grammatical gender in their native language.
Experience with a language that uses grammatical gender
appears to bias one’s representations of objects. This raises
a further question: what would happen if a person spoke two
languages that gave opposite grammatical genders to an
object? Would the discovery that grammatical genders can
vary lead people to discount grammatical gender biases?
Would people stick with the biases of whichever language
they learned first?
Experiment 2: Spanish-German Bilinguals
To investigate these questions, Experiment 2 tested people
who spoke both Spanish and German in the same similarity
task as described in Experiment 1.
Participants Thirty-six Spanish-German bilinguals who
were also fluent in English participated in the study on a
volunteer basis. Each participant had some experience with
both Spanish & German, and all were also proficient in
English. Participants had an average of 23.1 years of
experience with Spanish (ranging from 1 to 61 years), an
average of 26.1 years of experience with German (ranging
from 1 to 67 years), and an average of 21.5 years of
experience with English (ranging from 2 to 62 years). On a
scale of 1 (not fluent) to 5 (very fluent), participants rated
themselves on average 4.19 in Spanish, 4.27 in German, and
4.28 in English.
Materials and Procedures Materials and Procedures were
the same as in Experiment 1.
Results and Discussion For each subject, a similarity-rating
score was calculated by taking their average similarity
rating for gender-consistent pairs (relative to Spanish), and
subtracting their average similarity rating for gender-
inconsistent pairs. Since all objects used had opposite
genders in Spanish & German, a resulting positive score
would indicate a Spanish bias in similarity ratings, and a
negative score would indicate a German bias. Each subject’s
relative language skill was calculated by subtracting their
German-fluency score (provided on a scale of 1= not fluent
to 5=very fluent) from their Spanish-fluency score. Again, a
positive number indicated higher relative proficiency in
Spanish, and a negative number indicated higher relative
proficiency in German.
There was a significant positive correlation between
people’s relative proficiency in Spanish/German and their
biases in the similarity task, r=.40, N=36, p<.01. The more
relatively proficient a subject was in Spanish, the more
consistent with Spanish grammatical gender their similarity
ratings. And the more relatively proficient a subject was in
German, the more consistent with German grammatical
gender their similarity ratings. Subjects’ relative skill in the
two languages was the best predictor of their similarity
scores, faring better than whether one was born in a Spanish
or German speaking country (r=.325, N=36, p<.05), and
which of the two languages one has known for a longer
amount of time (r=.158, N=36, p=.178).
Experiment 3: Verbal Interference
We repeated Experiment 1 with the addition of a verbal
interference task, in order to rule out the hypothesis that the
effect is due to the subjects subvocally naming the objects.
Participants Seven Spanish-English bilinguals and seven
German-English bilinguals participated in the study in
exchange for payment. Both groups were highly proficient
in English and had an average of 17.04 years of experience
with English (19.57 years for German speakers, and 14.5
years for Spanish speakers). None of the Spanish speakers
spoke German and none of the German speakers spoke
Spanish. None of the subjects had a self-rated proficiency
greater than 2 (on a scale of 1 to 5, where 5 is fluent) in any
other languages that use grammatical gender.
Materials All of the similarity task materials were exactly
the same as in Experiment 1. In addition to the similarity
task, subjects performed a verbal shadowing task. A
computer played an audio-stream of randomly generated
English letters at a speed of 1 per second, and subjects were
asked to repeat each letter aloud as it was played.
Procedure The procedures were similar to Experiment 1
with the following changes. Subjects were asked to perform
the shadowing task described above the entire time that they
were rating similarities. Subjects’ verbal shadowing
performance was recorded by a tape-recorder. Just as in
Experiment 1, all of the participants were tested in English
with English instructions.
Results and Discussion Verbal interference did not change
Spanish and German speakers similarity ratings. Just as in
Experiment 1, Spanish and German speakers rated person—
object pairs more similar when the grammatical gender of
the object’s name in their native language was consistent
with the biological gender of the person in the comparison
(M=3.15) than when the two genders were inconsistent (M=
2.98), t= 2.20, df= 13, p<.05 (by subjects). This interaction
between person-gender and object-gender was also
confirmed in a 2 x2 repeated measures ANOVA (2 person
gender X 2 object-gender), F(1, 13)= 4.85, p<.05 (by
subjects). Because all of the objects chosen for this study
had opposite genders in Spanish and German, Spanish and
German speakers rated opposite pairs as more similar. This
was confirmed by a 3-way interaction of person-gender,
object-gender (relative to one of the languages), and native
language, F(1, 12)= 4.77, p<.05 (by subjects).
Comparing the results of Experiments 1 and 3, there was
no effect of the verbal interference task on the effect of
gender-consistency as confirmed by a lack of consistency by
interference interaction in a 2x2 repeated measures ANOVA
(2 (gender consistent or inconsistent) X 2 (interference
(Exp. 6) or no interference (Exp. 4)), F(1,67)=.075, p=.786
(by subjects), and F(1,26)=.073, p=.789 (by items). The
same analysis confirmed an overall main effect of
grammatical gender consistency both by subjects (F (1, 67)
= 8.18, p<.01) and by items (F(1,26) = 7.57, p<.02 across
the two experiments.
Finally, participants in the interference condition
performed at an average of 90% correct on the shadowing
task. This confirms that the failure of the verbal interference
manipulation to override the effect of consistency was not
due to the subjects’ failure to engage in the shadowing task.
These findings once again indicate that people’s thinking
about objects is influenced by the grammatical genders their
native language assigns to the objects’ names. A further
question is whether such differences in similarity can be
obtained just by differences in grammar, and without
concomitant cultural differences.
Experiments 4 and 5: Gumbuzi Similarity
Experiment 4 was designed to test whether grammatical
gender in a language can indeed exert a causal power over
thought without intermediary cultural factors. Native
English speakers were taught about the soupative/oosative
distinction in the fictional Gumbuzi language. Participants
were shown pictures of males and females along with many
inanimate objects and were taught which would be
considered soupative and which oosative in Gumbuzi. The
soupative/oosative distinction always corresponded to
biological gender (all females were in one category and all
males in the other) but also extended to inanimate objects. A
given participant might have learned that pans, forks,
pencils, ballerinas, and girls are soupative, while pots,
spoons, pens, giants, and boys are oosative.
After participants had mastered the oosative/soupative
distinction, they rated the similarity of each person-object
pair, much like the Spanish and German speakers in
Experiment 1. Experiment 5 adds verbal inteference
condition. We hypothesized that with or without
interference, same-gender pairs would be rated as more
Participants Twenty-two native English speakers
participated in the study in exchange for payment. Twelve
of these participated in the verbal interference version of the
task, while the remaining ten did not do the verbal
Materials A set of 20 pictures was constructed to include 8
pictures of people and 12 pictures of inanimate objects. The
inanimate objects were chosen in pairs such that the
members of each pair were quite similar to each other (e.g.,
fork and spoon, pen and pencil, bowl and cup, guitar and
violin, apple and pear, pot and pan). The members of each
pair were assigned to different grammatical categories (if
“pot” was oosative, “pan” would be soupative). This was
done so that participants had to pay attention to the
particulars of each object and couldn’t simply develop a
heuristic like “kitchen things are soupative” or fruits are
oosative”. Pears, forks, violins, pots, pens, and cups were
said to be oosative, and apples, spoons, guitars, pans,
pencils, and bowls were said to be soupative.
The pictures of people included 4 pictures of males (a
man, a boy, a giant, and a king), and 4 pictures of females (a
woman, a girl, a ballerina, and a bride). For half of the
participants the females were said to be oosative and the
males soupative, and for the other half it was the reverse.
This meant that each inanimate object was grouped with
females for half of the participants, and with males for the
other half. Overall, each participant learned to classify 20
pictures into two categories (oosative or soupative) with
each category containing 4 pictures of people of the same
gender and 6 pictures of inanimate objects.
Procedure Participants read the following instructions: “In
this study you will learn a bit about the Gumbuzi language.
In Gumbuzi, there are two different words for ‘the.’ For
example, in order to say ‘the chair’ you would saysou
chair,’ and in order to say ‘the table’ you would say ‘oos
table.’ This is called the oosative/soupative distinction.
Some nouns are always preceded by ‘sou’ and some are
always preceded by ‘oos.’Participants were then shown
pictures on a computer screen one at a time with each
picture accompanied by a label (a picture of a pear for
example would be accompanied by “oos pear”).
After they had seen all the items three times, they were
tested on how well they had learned the oosative/soupative
distinction. Participants were shown the pictures one at a
time and had to indicate whether each item would be
considered oosative or soupative in Gumbuzi by pressing
one of two keys on a keyboard. If they answered correctly,
the computer went on to the next item. If they answered
incorrectly, it beeped and waited for them to provide the
correct answer. Participants were tested until they could
answer all twenty items correctly in a row.
After they had learned the oosative/soupative distinction
perfectly, participants proceeded to the similarity-ratings
portion of the study. Just as the Spanish & German speakers
in Experiments 1-3, participants rated the similarity of every
possible person—object pair (a total of 96 comparisons) on
a scale of 1 (not similar) to 9 (very similar). The computer
generated a random order of pairs for each subject. Each
pair was kept on the screen until the participant made their
response by pressing one of the number keys on the
keyboard. One group of participants performed the
similarity-ratings task on its own, while another rated
similarities while at the same time performing the verbal
interference task described in Experiment 3.
Results and Discussion Just as the Spanish and German
speakers, participants in this study rated person—object
pairs more similar when they were consistent in gender
(M=4.43) than when the two genders were inconsistent
(M=3.79), F(1,11)=26.8, p<.001 (by items), F(1,20)=7.14,
p<.05 (by subjects). The effect was present equally for both
the verbal shadowing and the non-shadowing groups:
without verbal shadowing M=4.63 when the genders were
consistent and M=3.97 when they were inconsistent, t=3.24,
df=11, p<.01, and with verbal shadowing M=4.27 when the
genders were consistent and M=3.65 when they were
inconsistent, t=4.24, df=11, p<.001. There was no effect of
the shadowing task on the effect of consistency as
confirmed by a lack of consistency by shadowing
interaction in a 2x2 repeated measures ANOVA (2
(consistent or inconsistent) X 2 (shadowing or no
shadowing)), F(1,20)=.005, p=.942 (by subjects), and
F(1,11)=.019, p=.892 (by items). Finally, participants in the
shadowing condition performed at an average of 96%
correct on the shadowing task. This confirms that the failure
of the verbal shadowing manipulation to override the effect
of consistency was not due to the subjects’ failure to engage
in the shadowing task.
These results show that the effects of grammatical gender
on object representations can be produced in the absence of
culture, even under verbal interference.
Discussion Beyond demonstrating that learning linguistic
categories can affect people’s descriptions of objects or
similarity ratings,
General Discussion
These findings suggest that people’s ideas about the genders
of objects can indeed be influenced by the grammatical
genders assigned to those objects in a language. Further, this
effect can be produced just by grammatical differences and
in the absence of other cultural factors it is important to
consider how learning such categories can have this effect.
One possibility is that in order to make sense of the
grammatical categories they encounter in language (or in the
lab) people deliberately look for similarities between items
assigned to the same grammatical category. If a meaningful
and consistent set of similarities is discovered, these
similarities can then be stored (or perhaps the features that
are relevant to the similarity can be made more salient in the
representation), and this would explain both the increased
within-category similarity reported in this paper and the bias
in descriptions observed in the earlier studies (Boroditsky &
Schmidt, 2000; under review). This type of mechanism is
supported by recent findings suggesting that comparison
leads to an increase in similarity (so long as the items being
compared make it possible to discover meaningful
similarities) (Boroditsky, under review; see also Gentner &
Namy, 1999; Lowenstein & Gentner, 1998).
One interesting question is how learning what we’ve
called here a “grammatical categorycan change people’s
ideas about objects. One possibility is that grammatical
categories in language function just like other category
names (e.g., penguin, game, etc). By calling things by the
same name, or putting them in the same grammatical
category, languages may invite their speakers to (not
necessarily consciously) carry out comparisons that they
wouldn’t have otherwise carried out (or perhaps wouldn’t
have carried out as often or with the same goals in mind). In
the process of carrying out these comparisons, people may
discover meaningful similarities between objects. Any
discovered similarities may then be stored or highlighted in
the representations of the objects. There is no claim being
made here about the specialness of language in having this
effect. There are many ways to direct a person’s attention to
the similarities or differences of a pair of items, or to
classify a group of items into two categories language just
happens to be a popular and convenient medium for doing
So does all this mean that language affects thought? Or,
more precisely, does all this mean that thinking for speaking
a particular language can have an effect on how people
think even when not thinking for that same language?
The results reviewed and described in this paper
demonstrate that a grammatical distinction in language has
the power to bias people’s memory for and their
descriptions of objects and has an effect on people’s ratings
of similarity between pictures of objects. This is true even
though people perform tasks in a language different from
the one they learned the grammatical distinction in, perform
tasks involving no words (just pictures), and even despite
interference from a verbal shadowing task. Previous
evidence also suggests that the same grammatical distinction
affects people’s decision making (e.g., assigning voices to
animated characters), personification of nouns (as in the
Russian days of the week), and ratings of object
characteristics (e.g., potency). In short, speakers of different
languages behave differently in a wide range of cognitive
tasks in ways that are consistent with the grammatical
distinctions made in their languages.
But does all this evidence mean that language affects
thought? In particular, does it mean that linguistic categories
(e.g., a noun being grammatically feminine or masculine)
actually alter non-linguistic representations? Perhaps
linguistic categories simply get recruited covertly for all
these tasks, so even though speakers of different languages
may exhibit different patterns in behavior, linguistic and
non-linguistic representations remain truly separate, and
everybody’s non-linguistic representations are in fact the
This is an interesting possibility, and a difficult one to
rule out empirically. The fact that grammatical distinctions
learned in one language seem to have an effect even when a
task is performed in another language may favor the view
that grammatical knowledge actually plays a role in shaping
the underlying non-linguistic representation (hence the ease
of transfer between languages). Still, this kind of evidence
does not rule out the possibility that grammatical
information (even from the wrong language) is covertly
recruited in all sorts of tasks that don’t seem to require it. To
test this further, we attempted to disable people’s linguistic
faculties by asking them to shadow speech while they
performed the similarity-rating tasks described earlier. If
effects of grammatical gender had disappeared under these
verbal interference conditions, then we may have been able
to infer that grammatical categories hadn’t affected non-
linguistic representations.
Instead, it would seem that language affected thinking in
this case because people covertly invoked linguistic
representations in a set of seemingly non-linguistic tasks.
But it turned out that tying up the linguistic faculties had no
effect on the results (that effects of grammatical gender
were equally strong when subjects were under verbal
interference as when not). Can we now conclude that
grammatical gender definitely does affect people’s non-
linguistic representations? Perhaps the shadowing task
simply didn’t disable all of the aspects of language that
could have been covertly recruited for the task? Perhaps
some different, more complex verbal interference task
would have changed the results. Several other tasks could be
tried, but as long as the verbal interference doesn’t get rid of
the effect of language on thought, there will always be doubt
about whether or not all of the necessary linguistic faculties
were properly interfered with. There seems to be no sure
way to disable all linguistic processes (and this is not in
small part due to the difficulty in deciding on what counts as
linguistic and non-linguistic processing in the first place).
Fortunately, being able to discriminate between these two
possibilities is not necessary here. Most likely, both
possibilities are true to some degree. Regardless of the
answer, it appears that language plays an important role in
thinking. Whether people’s native language is covertly
involved in all manner of seemingly non-linguistic tasks
(even despite verbal interference, in tasks conducted entirely
in pictures, and in tasks conducted in other languages), or
whether aspects of grammar are able to influence non-
linguistic representations directly, it appears that (what we
colloquially call) thinking involves a collaboration between
many different linguistic and non-linguistic representations
and processes. This means that the private mental lives of
speakers of different languages may differ dramatically
and not only when they are thinking for speaking their
particular languages, but in all manner of cognitive tasks.
The findings presented in this paper suggest that people’s
thinking about objects can be influenced by aspects of
grammar that differ across languages. A series of studies
found effects of grammatical gender on people’s perceptions
of similarity between objects and people. This was true even
though the tasks were performed in English (a language
devoid of grammatical gender), even when the tasks were
non-linguistic (e.g., rating similarities between unlabeled
pictures), and even while subjects were engaged in a verbal
interference task. Finally, results showed that cross-
linguistic differences in thought can be produced just by
grammatical differences and in the absence of other cultural
factors. It is striking that even a fluke of grammar (the
nearly arbitrary assignment of a noun to be masculine or
feminine) can have an effect on how people think about
things in the world. Considering the many ways in which
languages differ, our findings suggest that the private mental
lives of people who speak different languages may differ
much more than previously thought.
This research was funded by an NSF Graduate Research
Fellowship to Lera Boroditsky. Partial support was also
provided by NIMH research grant MH-47575 to Gordon
Bower. The authors would like to thank Michael Ramscar,
Herbert H. Clark, Eve Clark, Barbara Tversky, Gordon
Bower, and Steven Pinker for helpful comments and
insightful discussions of this research. We would also like to
thank Lauren Schmidt for her foundational contributions to
this research, and Jill M. Schmidt who generously
contributed most of the pictures used in these studies.
Chalnick, A., & Billman, D. (1988). Unsupervised learning
of correlational structure. Proceedings of the Tenth
Annual Conference of the Cognitive Science Society (pp.
510-516). Hillsdale, NJ: Lawrence Erlbaum Associates.
Feigenbaum, E. A. (1963). The simulation of verbal
learning behavior. In E. A. Feigenbaum & J. Feldman
(Eds.), Computers and thought. New York: McGraw-Hill.
Boroditsky, L. (in preparation). The role of comparison in
the development of similarity.
Boroditsky, L. (2001). Does Language Shape Thought?
Mandarin and English speakers conceptions of time.
Cognitive Psychology, 43(1).
Boroditsky, L. (2000). Metaphoric structuring:
Understanding time through spatial metaphors. Cognition,
Boroditsky, L. & Schmidt, L. (2000). Sex, Syntax, and
Semantics. Proceedings of the 22nd Annual Meeting of the
Cognitive Science Society.
Braine, M. (1987). What is learned in acquiring word
classes - a step toward an acquisition theory. In B.
MacWhinney (Ed.), Mechanisms of language acquisition.
Hillsdale, NJ: Erlbaum.
Fodor, I. (1959). The origin of grammatical gender I.
Lingua, 8, 1.
Caramazza, A., & Miozzo, M. (1997). The relation between
syntactic and phonological knowledge in lexical access:
Evidence from the "tip-of-the-tongue" phenomenon.
Cognition, 64(3).
Gentner, D., & Namy, L. (1999). Comparison in the
development of categories. Cognitive Development, 14,
Loewenstein, J., & Gentner, D. (1998). Relational language
facilitates analogy in children. Proceedings of the
Twentieth Annual Conference of the Cognitive Science
Society, 615-620. Mahwah, NJ: Lawrence Erlbaum
Sapir, E. (1921). Language, New York, NY: Harcourt,
Brace, and World.
Slobin, D. (1996). From “thought and language” to
“thinking for speaking.” In J. Gumperz & S. Levinson
(Eds.), Rethinking linguistic relativity. Cambridge, MA:
Cambridge University Press..
Vigliocco, G., Antonini, T., & Garrett, M.F. (1997).
Grammatical gender is on the tip of Italian tongues.
Psychological-Science, 8(4).
... This further supports the idea that grammatical gender can affect how an object is conceptualized, independent of language (the study was conducted in English and yet the adjectives were conditioned by the grammatical gender in the participants' first language). It is plausible this is because cultural artefacts (like personification in fairy tales) act as an intermediary, affecting the concept, but Boroditsky and Phillips (2002) conducted an experiment on English natives using an artificial language with a gender distinction that extended to non-animate referents and found that this too affected how objects were described, with more feminine adjectives being used to describe the objects belonging to the same noun class as females. ...
Full-text available
Undergraduate dissertation focusing on the resolution of conflicting grammatical and meaningful gender, particularly in situations of metonymy.
... Many studies on bilinguals (e.g., Boroditsky et al., 2003;Phillips and Boroditsky, 2003), as well as on monolinguals (e.g., Imai et al., 2014;Saalbach et al., 2012;Sera et al., 2002), similarly point to the impact of grammatical gender on mental representations. In other words, depending on the language at use, readers and speakers are pointed to the gender features associated with the grammatical category. ...
... Conversely, stereotypically female qualities are used to describe the smaller cars: agility, beauty, sensuality, inter alia: Sirion: the most agile (F) in town; Hyundai, ignites (F) the imagination; Opel, a naughty (F) German. These findings reverberate well with studies that showed a congruence between grammatical gender and perception of femininity/masculinity of objects or animals (Haertlé, 2017;Phillips & Boroditsky, 2003). ...
Full-text available
In urban centers, many residents commute to work and are exposed to visual input displayed on cars around them. License plates become a space utilized by advertisers to display advertising messages. This is an instance of the presence of language in the public space, or the linguistic landscape of an area. The purpose of the present study is to analyze linguistic and visual elements of license plates in the Israeli linguistic landscape. The study positions the license plates as a medium that conveys messages reflecting Israeli society and culture. We focus on the advertising texts on license plate frames and examine the messages they carry about the car and the car owner. Through a content analysis of such texts, findings point to varied linguistic-visual elements in the license plates from a linguistic landscape perspective, such as the presence of languages and the flag, including the empowerment of the car company and the car itself, the driving experience, driver’s superiority and dream realization, and the personal relationships between the driver and the car. We conclude that today cars are not only vehicles taking us from place to place, but are also associated with power, emotion, freedom, superiority, and self-expression.
Many languages assign nouns to grammatical gender categories (e.g., masculine and feminine), and inanimate objects often have different genders in different languages. In a seminal study, Phillips and Boroditsky (2003) provided evidence that such “quirks of grammar” influence how people conceptualize objects. Spanish and German speakers judged person-object picture pairs as more similar when their biological and grammatical genders matched than when they did not, and English speakers showed the same pattern of similarity judgments after learning gender-like categories. These widely cited findings were instrumental in vindicating the Whorfian hypothesis that language shapes thought, yet neither the original study nor any direct replications have appeared in a peer-reviewed journal. To examine the reliability of Phillips and Boroditsky’s findings, we conducted a high-powered replication of two of their key experiments (total N = 375). Our results only partially replicated the original findings: Spanish and German speakers’ similarity judgments exhibited no effect of grammatical gender when accounting for key sources of error variance, but English speakers trained on gender-like categories rated same-gender pairs more similar than different-gender pairs. These results provide insight into the contexts in which grammatical gender effects occur and the mechanisms driving them.
Full-text available
There is empirical evidence in different languages on how the computation of gender morphology during psycholinguistic processing affects the construction of sex-generic representations. However, there are few experimental studies in Spanish and there is no empirical evidence that analyzes the psycholinguistic processing of morphological innovations used as non-binary forms (-x; -e) in contrast to the generic masculine variant (-o). To analyze this phenomenon, we designed a sentence comprehension task. We registered reading times, precision and response times. The results show the specialization of non-binary forms as generic morphological variants, as opposed to the generic masculine. The non-binary forms consistently elicited a reference to mixed groups of people and the response times indicated that these morphological variants do not carry a higher processing cost than the generic masculine. Contrary to what classical grammatical approaches propose, the generic masculine does not function in all cases as generic and its ability to refer to groups of people without uniform gender seems to be modulated by the stereotipicality of the role names.
Full-text available
Across many languages, pronouns are the most frequently produced referring expressions. We examined whether and how speakers avoid referential ambiguity that arises when the gender of a pronoun is compatible with more than one entity in the context in French. Experiment 1 showed that speakers use fewer pronouns when human referents have the same gender than when they had different genders, but grammatical gender congruence between inanimate referents did not result in fewer pronouns. Experiment 2 showed that semantic similarity between non-human referents can enhance the likelihood that speakers avoid grammatical-gender ambiguous pronouns. Experiment 3 pitched grammatical gender ambiguity avoidance against the referents' competition in the non-linguistic context, showing that when speakers can base their pronoun choice on non-linguistic competition, they ignore the pronoun's grammatical gender ambiguity even when the referents are semantically related. The results thus indicated that speakers preferentially produce referring expressions based on non-linguistic information; they are more likely to be affected by the referents' non-linguistic similarity than by the linguistic ambiguity of a pronoun.
Full-text available
We use large-scale corpora in six different gendered languages, along with tools from NLP and information theory, to test whether there is a relationship between the grammatical genders of inanimate nouns and the adjectives used to describe those nouns. For all six languages, we find that there is a statistically significant relationship. We also find that there are statistically significant relationships between the grammatical genders of inanimate nouns and the verbs that take those nouns as direct objects, as indirect objects, and as subjects. We defer deeper investigation of these relationships for future work.
Full-text available
The influence of grammatical gender on cognitive processes is an important issue in contemporary psycholinguistics and language psychology, particularly in research concerning the relations between grammar and semantics. The extent of this effect is dependent on a given language’s gender system and its grammatical specifics. The aim of the presented research was to investigate grammatical gender effects in Polish – a Slavic language with three singular and two plural grammatical genders. In Experiment 1, triadic similarity judgments were used, and it turned out that the grammatical gender of nouns influenced perceived similarity of words in case of animals, but not inanimate objects or abstract concepts. In Experiment 2 we used a modified Implicit Association Test; results suggest that grammatical gender seems to be of implicit nature, as grammatical gender consistency influenced reaction times and the number of classification errors. In Experiment 3 participants assigned male and female voices to animals and inanimate objects, which were presented either as words or as pictures. Grammatical gender effects occurred for both animate and inanimate objects and were similar for verbal and visual stimuli. It turned out that in the Polish language the influence of grammatical gender may occur on the lexicosemantic level and the conceptual level, and concerns both animate and inanimate objects. Results are discussed in context of the similarity and gender and the sex and gender hypotheses.
Full-text available
This study investigates whether there is a relation between how motion is linguistically expressed and how it is conceptualised. To do this, native speakers of two languages that differ typologically in how they encode telic motion (English and Spanish) are compared in both a verbal and a non-verbal experiment. The preferred non-verbal methods to test the linguistic relativity hypothesis in this domain have so far been recognition memory and binary judgments. This study questions the experimental validity of these approaches and implements an alternative method which combines similarity ratings with a verbal interference manipulation. The results reported here constitute evidence against linguistic relativity and in support of cognitive universalism.
Full-text available
Abstract One important function of language is to name,relations. Pre- school children performed ,a simple ,mapping ,task with and without hearing spatial prepositions calling attention to key relations. Children at 44 months,were successful only if they were in the language condition. By 49 months, children were competent on the task regardless of condition, although there were still benefits of language. These results suggests that re- lational language can therefore be an important tool for high- lighting relational commonalities,children may,otherwise fail to use.
To correctly produce words, speakers must have access to three broad classes of information lexical semantics, syntax, and sound structure The relevant information must be organized in ways that permit rapid and accurate retrieval of specific lexical targets Current models of language production do this by a two-stage process The first stage incorporates lexical meanings and syntax, and the second, sound structure We used studies of the tip-of-the-tongue phenomenon (i e, the condition in which a speaker cannot produce a well-known word) to evaluate this organization and in so doing, we provide the first clear experimental evidence for a lexical stage that includes syntax and is distinct from both sound structure and the conceptual correlates of syntactic features.
An information processing model of elementary human symbolic learning is given a precise statement as a computer program, called Elementary Perceiver and Memorizer (EPAM). The program simulates the behavior of subjects in experiments involving the rote learning of nonsense syllables. A discrimination net which grows is the basis of EPAM's associative memory. Fundamental information processes include processes for discrimination, discrimination learning, memorization, association using cues, and response retrieval with cues. Many well-known phenomena of rote learning are to be found in EPAM's experimental behavior, including some rather complex forgetting phenomena. EPAM is programmed in Information Processing Language V. H. A. Simon has described some current research in the simulation of human higher mental processes and has discussed some of the techniques and problems which have emerged from this research. The purpose of this paper is to place these general issues in the context of a particular problem by describing in detail a simulation of elementary human symbolic learning processes. The information processing model of mental functions employed is realized by a computer program called Elementary Perceiver and Memorizer (EPAM). The EPAM program is the precise statement of an information processing theory of verbal learning that provides an alternative to other verbal learning theories which have been proposed. It is the result of an attempt to state quite precisely a parsimonious and plausible mechanism sufficient to account for the rote learning of nonsense syllables. The critical evaluation of EPAM must ultimately depend not upon the interest which it may have as a learning machine, but upon its ability to explain and predict the phenomena of verbal learning. I should like to preface my discussion of the simulation of verbal learning with some brief remarks about the class of information processing models of which EPAM is a member. a. These are models of mental processes, not brain hardware. They are psychological models of mental function. No physiological or neurological assumptions are made, nor is any attempt made to explain information processes in terms of more elementary neural processes. b. These models conceive of the brain as an information processor with sense organs as input channels, effector organs as output devices, and with internal programs for testing, comparing, analyzing, rearranging, and storing information. c. The central processing mechanism is assumed to be serial; i.e., capable of doing only one (or a very few) things at a time. d. These models use as a basic unit the information symbol ; i.e., a pattern of bits which is assumed to be the brain's internal representation of environmental data. e. These models are essentially deterministic , not probabilistic. Random variables play no fundamental role in them.
Recent research on children's word learning has led to a paradox. Although word learning appears to be a deep source of insight into conceptual knowledge for children, preschoolers often categorize objects on the basis of shallow perceptual features such as shape. The current studies seek to resolve this discrepancy. We suggest that comparing multiple instances of a category enables children to extract deeper relational commonalities among category members. We examine 4-year-olds' categorization behaviors when asked to select a match for a target object (e.g., an apple) between a perceptually similar, out-of-kind object (e.g., a balloon) and a perceptually different category match (e.g., a banana). Children who learn a novel word as a label for multiple instances of the category are more likely to select the category match over the perceptual match. Children who learn a label for only one instance are equally likely to select either alternative. This effect is present even when individual target instances are more perceptually similar to the perceptual choice than to the category choice. We conclude that structural alignment processes may be important in the development of category understanding.
The relation between access to the syntactic and to the phonological features of words in lexical access is investigated in two experiments. Italian speakers were asked to provide the gender and partial phonological information of known nouns they could not produce at that moment, words that they felt were at the tip-of-the-tongue (TOT). In both experiments, subjects were able to provide information about the word they could not produce with better-than-chance accuracy. This was true not only for phonological information such as the initial phoneme of the word but also for the word's gender--a purely syntactic feature of nouns. However, analyses of the correlation between correct retrieval of the gender and the initial phoneme failed to reveal a positive relationship. This result is inconsistent with theories of lexical access the interpose two lexical nodes, lemma and lexeme nodes, between a word's semantic and phonological content. A model of lexical access that does not postulate the lemma/lexeme distinction is briefly discussed.
The present paper evaluates the claim that abstract conceptual domains are structured through metaphorical mappings from domains grounded directly in experience. In particular, the paper asks whether the abstract domain of time gets its relational structure from the more concrete domain of space. Relational similarities between space and time are outlined along with several explanations of how these similarities may have arisen. Three experiments designed to distinguish between these explanations are described. The results indicate that (1) the domains of space and time do share conceptual structure, (2) spatial relational information is just as useful for thinking about time as temporal information, and (3) with frequent use, mappings between space and time come to be stored in the domain of time and so thinking about time does not necessarily require access to spatial schemas. These findings provide some of the first empirical evidence for Metaphoric Structuring. It appears that abstract domains such as time are indeed shaped by metaphorical mappings from more concrete and experiential domains such as space.
Does the language you speak affect how you think about the world? This question is taken up in three experiments. English and Mandarin talk about time differently--English predominantly talks about time as if it were horizontal, while Mandarin also commonly describes time as vertical. This difference between the two languages is reflected in the way their speakers think about time. In one study, Mandarin speakers tended to think about time vertically even when they were thinking for English (Mandarin speakers were faster to confirm that March comes earlier than April if they had just seen a vertical array of objects than if they had just seen a horizontal array, and the reverse was true for English speakers). Another study showed that the extent to which Mandarin-English bilinguals think about time vertically is related to how old they were when they first began to learn English. In another experiment native English speakers were taught to talk about time using vertical spatial terms in a way similar to Mandarin. On a subsequent test, this group of English speakers showed the same bias to think about time vertically as was observed with Mandarin speakers. It is concluded that (1) language is a powerful tool in shaping thought about abstract domains and (2) one's native language plays an important role in shaping habitual thought (e.g., how one tends to think about time) but does not entirely determine one's thinking in the strong Whorfian sense.