ArticlePDF Available

Abstract and Figures

Gygax, Gabriel, Sarrasin, Oakhill, and Garnham (2008) showed that readers form a mental representation of gender that is based on grammatical gender in French and German (i.e., masculine supposedly interpretable as a generic form) but is based on stereotypical information in English. In this study, a modification of their stimulus material was used to examine the additional potential influence of pronouns. Across the three languages, pronouns differ in their grammatical gender marking: The English they is gender neutral, the French ils is masculine, and the German sie, although interpretable as generic, is morphologically feminine. Including a later pronominal reference to a group of people introduced by a plural role name significantly altered the masculine role name's grammatical influence only in German, suggesting that grammatical cues that match (as in French) do not have a cumulative impact on the gender representation, whereas grammatical cues that mismatch (as in German) do counteract one another. These effects indicate that subtle morphological relations between forms actually used in a sentence and other forms have an immediate impact on language processing, although information about the other forms is not necessary for comprehension and may, in some cases, be detrimental to it.
Content may be subject to copyright.
This article was downloaded by: [Unviersité de Fribourg]
On: 18 December 2013, At: 06:28
Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954
Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH,
UK
Discourse Processes
Publication details, including instructions for
authors and subscription information:
http://www.tandfonline.com/loi/hdsp20
Gender Representation in
Different Languages and
Grammatical Marking on
Pronouns: When Beauticians,
Musicians, and Mechanics
Remain Men
Alan Garnham a , Ute Gabriel b , Oriane Sarrasin c ,
Pascal Gygax c & Jane Oakhill a
a School of Psychology, University of Sussex ,
Brighton , United Kingdom
b Department of Psychology , Norwegian University
of Science and Technology , Trondheim , Norway
c Department of Psychology , University of Fribourg ,
Fribourg , Switzerland
Accepted author version posted online: 08 May
2012.Published online: 06 Aug 2012.
To cite this article: Alan Garnham , Ute Gabriel , Oriane Sarrasin , Pascal Gygax &
Jane Oakhill (2012) Gender Representation in Different Languages and Grammatical
Marking on Pronouns: When Beauticians, Musicians, and Mechanics Remain Men,
Discourse Processes, 49:6, 481-500, DOI: 10.1080/0163853X.2012.688184
To link to this article: http://dx.doi.org/10.1080/0163853X.2012.688184
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the
information (the “Content”) contained in the publications on our platform.
However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness,
or suitability for any purpose of the Content. Any opinions and views
expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the
Content should not be relied upon and should be independently verified with
primary sources of information. Taylor and Francis shall not be liable for any
losses, actions, claims, proceedings, demands, costs, expenses, damages,
and other liabilities whatsoever or howsoever caused arising directly or
indirectly in connection with, in relation to or arising out of the use of the
Content.
This article may be used for research, teaching, and private study purposes.
Any substantial or systematic reproduction, redistribution, reselling, loan,
sub-licensing, systematic supply, or distribution in any form to anyone is
expressly forbidden. Terms & Conditions of access and use can be found at
http://www.tandfonline.com/page/terms-and-conditions
Downloaded by [Unviersité de Fribourg] at 06:28 18 December 2013
Discourse Processes, 49:481–500, 2012
Copyright © Taylor & Francis Group, LLC
ISSN: 0163-853X print/1532-6950 online
DOI: 10.1080/0163853X.2012.688184
Gender Representation in Different
Languages and Grammatical Marking
on Pronouns: When Beauticians,
Musicians, and Mechanics Remain Men
Alan Garnham
School of Psychology
University of Sussex, Brighton, United Kingdom
Ute Gabriel
Department of Psychology
Norwegian University of Science and Technology, Trondheim,Norway
Oriane Sarrasin and Pascal Gygax
Department of Psychology
University of Fribourg, Fribourg, Switzerland
Jane Oakhill
School of Psychology
University of Sussex, Brighton, United Kingdom
Gygax, Gabriel, Sarrasin, Oakhill, and Garnham (2008) showed that readers form
amentalrepresentationofgenderthatisbasedongrammatical gender in French
and German (i.e., masculine supposedly interpretable as a generic form) but is
based on stereotypical information in English. In this study, a modificatio n of
their stimulus material was used to examine the additional potential influence of
pronouns. Across the three languages, pronouns differ in their grammatical gender
marking: The English they is gender neutral, the French ils is masculine, and
Correspondence concerning this article should be addressedtoAlanGarnham,SchoolofPsychol-
ogy, University of Sussex, Brighton, BN1 9QH, United Kingdom. E-mail: a.garnham@sussex.ac.uk
481
Downloaded by [Unviersité de Fribourg] at 06:28 18 December 2013
482 GARNHAM ET AL.
the German sie,althoughinterpretableasgeneric,ismorphologicallyfeminine.
Including a later pronominal reference to a group of people introduced by a plural
role name significantly altered the masculine role name’s grammatical influence
only in German, suggesting that grammatical cues that match (as in French) do not
have a cumulative impact on the gender representation, whereas grammatical cues
that mismatch (as in German) do counteract one another. Theseeffectsindicatethat
subtle morphological relations between forms actually usedinasentenceandother
forms have an immediate impact on language processing, although information
about the other forms is not necessary for comprehension and may, in some cases,
be detrimental to it.
Areaderofthesentence“Thesingerhadcaughtacold”cannotbe sure whether
it is about a man or a woman. Nevertheless, research has shown that people
elaborate their mental representation of the singer to include gender. This rep-
resentation is part of a mental model of the text—more specifically, part of
the situation level of the mental model, which contains information about the
people, settings, actions, and events described either explicitly or implied by the
text (Garnham & Oakhill, 1996; Zwaan & Radvansky, 1998).
In English, gender representations from role names (i.e., any name that
incorporates features used to describe a person or a group of persons, such as
names indicating hobbies or pastimes—e.g., soccer fan—or occupations—e.g.,
dentist, actor, or student) are influenced by gender stereotypes (e.g., Carreiras,
Garnham, Oakhill, & Cain, 1996; Duffy & Keir, 2004; Garnham, Oakhill,
&Reynolds,2002;Kennison&Trofe,2003;Sturt,2003).Usingasentence
evaluation paradigm (Tanenhaus & Carlson, 1990) in which participants have
to judge whether a sentence is a sensible continuation of the preceding text,
Garnham et al., for example, found that participants had mosttroublewith,and
took longer to respond to, sentences that were incongruent with the stereotypical
gender of the role names in the preceding text. In a later study, Oakhil l, Garnham,
and Reynolds (2005) asked participants to judge whether two words (e.g.,
nurse and uncle)couldapplytothesameperson.Theyfoundthatwhenthe
experimenters attempted to suppress participants’ use of such information (e.g.,
by reminding them that many professions that were traditionally performed by
one gender are now performed by both), mental representations of gender were
still stereotyped, although the effects were reduced.
Whereas English readers rely on stereotypical gender, readers of correspond-
ing material in a language in which nouns carry grammatical gender might also
use grammatical information to infer the protagonist’s gender. In grammatically
gender-marked languages, such as German, French, or Spanish, the gender of a
character in a text is often explicitly represented by the form of the determiner
and by the morphologically feminine or masculine form of the noun. Thus, in
German, the sentence “Die Sängerin (Dfeminine form) hatte sich erkältet
Downloaded by [Unviersité de Fribourg] at 06:28 18 December 2013
GENDER REPRESENTATION AND PRONOUNS 483
‘The (female) singer had caught a cold’ unequivocally signifies that a woman
is referred to. Carreiras et al. (1996, see Experiments 2–4) showed that, in
Spanish, a clash between a determiner’s grammatical gender and the stereotyped
gender of a noun (e.g., la carpintera ‘the [female] carpenter’ [male stereotype,
female noun form]; cf. el carpintero ‘the [male] carpenter’) had an immediate
effect, even if the noun was not itself morphologically marked for gender (e.g.,
la futbolista ‘the [female] footballer’ [male stereotype form used for both males
and females]; cf. el futbolista ‘the [male] footballer’). Furthermore, a later
pronoun that mismatched the role noun’s stereotypical gender did not cause any
additional problems (e.g., La carpintero ::: Ella ::: ‘the (female) carpenter :::
She :::’). In many ways, this effect is to be expected, given the close(spatial
and grammatical) relations between noun and determiner and the fact that the
two together form an expression that refers to a single person.
In the Carreiras et al. (1996) materials, the noun phrases were intended to
refer to specific individuals, both for masculine and feminine forms. However,
there is a complication in the use of feminine and masculine forms in many
gender-marked languages. Whereas the use of feminine forms of role names
is unequivocal, the same is not true for the masculine forms, as those forms
are used in two different ways: specifically, referring to male persons; and
generically, referring to a group of persons of both genders,orreferringto
apersonoragroupofpersonsofunknowngendersandincontexts where
the gender of a person is irrelevant. This generic use of the masculine is
governed by explicit grammatical rules (Académie Française, 2002; Baudino,
2001). However, recent versions of the influential German grammar (Duden,
2009) refer to the generic merely as one of two Gebrauchsweisen (‘uses’) of
masculine nouns. Nevertheless, although there are guidelines on how to avoid
the use of the masculine-only in official announcements, the masculine is still
commonly used as generic in spoken, as well as in written, language. The basis
on which readers decide whether a masculine form is intended as generic or
specific remains unclear. However, generic uses are more common in the plural,
and the issue of possible generic interpretation did not arise in the materials
used by Carreiras et al.
Empirical research on the use of the masculine-as-generic (GM) in German
(for a review, see Braun, Sczesny, & Stahlberg, 2005), French(e.g.,Brauer&
Landry, 2008; Chatard, Guimond, & Martinot, 2005; Gygax & Gabriel, 2008),
and Norwegian (e.g., Gabriel, 2008; Gabriel & Gygax, 2008) indicates that
the use of the masculine evokes concepts of men, thus eliminating women as
potential referents in what should be generic uses (for a review, see Stahlberg,
Braun, Irmen, & Sczesny, 2007). These findings support the idea that readers of
gender-marked languages tend to interpret masculine forms as specific, at least
in situations where no other linguistic or non-linguistic information suggests
otherwise.
Downloaded by [Unviersité de Fribourg] at 06:28 18 December 2013
484 GARNHAM ET AL.
In a language such as German, the mental representation of gender derived
from the use of masculine role names is also affected by stereotype information
(Braun, Gottburgsen, Sczesny, & Stahlberg, 1998; Irmen & Roßberg, 2004;
Rothmund & Scheele, 2004), as in English, and supplemental grammatical
information (i.e., grammatical markings on pronouns and determiners: Rother-
mund, 1998). Rothermund investigated the mental representation of singular
and plural GM phrases in short texts that also included co-referential pronouns.
Surprisingly, the authors found a male bias for the interpretation of singular GM
phrases but a female bias for the plural GM phrases. Rothermund suggested
an influence of the plural determiner (i.e., die)andpronoun(i.e.,sie), both of
which have the same form as the feminine singular.
Gygax, Gabriel, Sarrasin, Oakhill, and Garnham (2008) investigated the inter-
play of gender stereotypicality and grammatical form (nounsanddeterminers)in
the representation of gender in English, French, and German.Usingrolenames
from a questionnaire study by Gabriel, Gygax, Sarrasin, Garnham, and Oakhill
(2008), also used in this research (and shown in Table 1), Gygax et al. asked
participants to read pairs of sentences in which the first sentence included a
role name as the subject (e.g., “The spies came out of the meeting room”), and
the second sentence contained explicit information about the character’s gender
(e.g., “It was obvious that one of the women was really angry”). Participants
had to decide whether the second sentence was a sensible continuation of the
first (Tanenhaus & Carlson [1990] argued that this task is specially suited to the
study of anaphoric processing). The sentences were identical in meaning in each
of the three languages, but in German and French, the role names were in the
masculine form, which, according to grammatical rules, should be interpreted
as a generic.
The results showed that, in English, participants’ gender representations of
the role names were in line with the role names’ stereotypicality. Participants
responded “yes” more often when the role name’s stereotypicality matched
the gender of the character in the second sentence (e.g., nurses followed by
women). In French and German, however, the representations were equally
male biased across all stereotypicality conditions. Participants responded “yes”
more often when the characters were men, independent of stereotypicality.
Participants’ answers in French and German were, therefore,stronglyinuenced
by the grammatical form of the noun, but there was no support for an influence
either of gender stereotypicality or of supplemental grammatical cues. Based
on Rothermund (1998), Gygax et al. (2008) had initially hypothesized that the
plural determiner in German (die), which is morphologically identical to the
feminine singular determiner, would counteract a male bias introduced by the
role name. It did not.
The role of gender marking on pronouns, particularly following referential
noun phrases that might be interpreted either generically orspecically,has
Downloaded by [Unviersité de Fribourg] at 06:28 18 December 2013
GENDER REPRESENTATION AND PRONOUNS 485
TABLE 1
Role Names Chosen From Gabriel, Gygax, Sarrasin, Garnham, and Oakhill (2008),
Along With the Proportion of Men Evaluated by Each Language Participant Group
English % German % French %
Male Stereotypes
Spies 73 Spione 67 Espions 74
Golfers 73 Golfspieler 68 Golfeurs 73
Politicians 71 Politiker 69 Politiciens 72
Police officers 63 Polizisten 69 Policiers 70
Statisticians 70 Statistiker 72 Statisticiens 74
Bosses 62 Arbeitgeber 72 Patrons 74
Computer specialists 70 Informatiker 79 Informaticiens 67
Surgeons 62 Chirurgen 75 Chirurgiens 75
Technic i a n s 72 Techniker 78 Te c h n i c iens 7 5
Engineers 78 Ingenieure 78 Ingénieurs 74
Physics students 56 Physikstudenten 81 Etudiants en physique 67
Pilots 70 Flieger 76 Aviateurs 74
M68 74 72
Neutral Stereotypes
Singers 53 Sänger 45 Chanteurs 48
Pedestrians 49 Spaziergänger 46 Promeneurs 52
Cinema goers 51 Kinobesucher 49 Spectateurs de cinéma 50
Concert goers 47 Konzert-Zuhörer 47 Auditeurs de concert 51
Schoolchildren 53 Schüler 48 Ecoliers 53
Spectators 55 Zuschauer 41 Spectateurs 51
Neighbors 50 Nachbarn 50 Voisins 50
Swimmers 50 Schwimmer 50 Nageurs 50
Tennis p l a y e rs 53 Tennis s p ieler 52 Jo u e u r s d e t e n n is 54
Authors 48 Autoren 52 Auteurs 54
Musicians 54 Musiker 50 Musiciens 59
Skiers 55 Skifahrer 53 Skieurs 55
M52 49 52
Female Stereotypes
Beauticians 29 Kosmetiker 11 Esthéticiens 18
Birth attendants 29 Geburtshelfer 11 Assistants maternels 18
Fortune tellers 32 Wahrsager 24 Diseurs de bonne aventure 28
Cashiers 39 Kassierer 27 Caissiers 24
Nurses 30 Krankenpfleger 24 Infirmiers 30
Hairdressers 48 Coiffeure 21 Coiffeurs 38
Psychology students 38 Psychologiestudenten 25 Etudiants en psychologie 33
Dieticians 39 Diätberater 27 Diététiciens 37
Dressmakers 43 Schneider/Näher 23 Couturiers 40
Dancers 32 Tänzer 33 Danseurs 29
Sales assistants 34 Verkäufer 33 Vendeurs 37
Social workers 29 Sozialarbeiter 41 Assistants sociaux 33
M35 24 30
Downloaded by [Unviersité de Fribourg] at 06:28 18 December 2013
486 GARNHAM ET AL.
not been properly investigated. One type of theory would suggest that gender
marking on pronouns is used in co-reference processing, but providing a match is
found, does not have further effects on intermediate to long-term representations
of the text, which might simply depend on the reactivated representation of the
antecedent phrase. A different type of account suggests thatsubtleaspectsofthe
gender information carried by the pronoun (e.g., whether itsformisidentical
to other pronominal forms in the language, as in German, wheresie is both
plural and feminine singular) can affect the (reactivated) representation of the
referent originally introduced by the role noun and its determiner. Note that,
like stereotype information, information about the relation between an actually
occurring form and other forms in the language is not only unnecessary for
comprehension, but it might also interfere with comprehension (e.g., if a German
plural sie is used to refer to a group of all males). As Rothermund’s (1998)
experiment could not differentiate between the inuence of the determiners and
the influence of the pronouns, one possible explanation for Gygax et al.’s (2008)
failure to replicate Rothermund’s findings is that, in the original experiment, the
effect was triggered by the pronominal anaphors and not by thedeterminers.
In the experiment presented here, we further investigated this issue by eval-
uating the extent to which adding pronouns can alter readers’representationsof
gender. More specifically, we drew on the fact that the three languages differ
not only in the gender markedness of the plural determiners, but also in their
pronoun systems.
In English, as well as in French, the plural definite determiner is gender
neutral (the and les), whereas in German, the plural definite determiner (die)
is morphologically identical with the feminine singular determiner (e.g., sin-
gular feminine: die Wissenschaftlerin; plural masculine: die Wissenschaftler;
plural feminine: die Wissenschaftlerinnen). Furthermore,inEnglish,theneutral
pronoun they is used to refer to a group of only women, only men, or to a
mixed group. In French, different plural pronouns exist to describe a group
of only women (elles)andagroupofonlymen(ils), but it is the masculine
plural pronoun that is used as to refer to a mixed group. In German, reference
to a mixed group is entirely different: The generic plural pronoun (sie)is
morphologically identical to the singular feminine (sie), but different from the
masculine singular (er). To summarize, when reference is made to a mixed
group, in English the pronoun is gender neutral, in French it is masculine, and
in German it is identical to the feminine singular.
Because referential pronouns are gender marked, we hypothesized that in
German and French they might have different implications forreaders’mental
representations of gender. In line with the notion that subtle aspects of morpho-
logical marking on pronouns (i.e., the relation between the form used and other
forms in the same language) affect gender representation, wehypothesizedthe
following:
Downloaded by [Unviersité de Fribourg] at 06:28 18 December 2013
GENDER REPRESENTATION AND PRONOUNS 487
H1: In English, the mental representation of gender should remain biased,
as in Gygax et al. (2008), by stereotyped information only.
H2: In French, the male bias found in Gygax et al. (2008) shouldbemain-
tained, and possibly enhanced, by the additional generic useofthe
masculine pronouns.
H3: In German, however, the male bias found in Gygax et al. (2008) should
be weakened by the use of the generic pronouns that are morphologically
identical to the feminine singular ones.
Alternatively, if these subtle morphological relations do not affect gender
representations, we would expect similar results in this study to those of Gygax
et al. (2008). Both outcomes are broadly compatible with the mental models
framework, which is primarily concerned with eventual representations of con-
tent. However, it is clear that the construction of the correct mental model does
not require the use of the subtle morphological relations referred to earlier,
so a result indicating that this information is used shows that considerations
other than the construction of mental models determine the architecture of the
language processing system.
METHOD
Participants
English sample. Thirty-six students (5 men and 31 women1)fromthe
University of Sussex took part in this experiment (age: MD22.03, SD D3.78).
Each participant was paid four pounds or received course credits.
French sample. Thirty-four students (2 men and 32 women) from the
University of Fribourg took part in this experiment (age: MD22.30, SD D
4.67). The participants received course credits or voluntarily took part in the
study.
German sample. Thirty-six students from the University of Bern (all
women) took part in this experiment (age: MD22.43, SD D3.54). The
participants received course credits.
1As Gygax, Gabriel, Sarrasin, Oakhill, and Garnham (2008) andothershavefoundnogender
of respondent differences in this kind of reading/judgment task and because we had access to many
more female participants than male participants, we decidedtonotbalancethesample.
Downloaded by [Unviersité de Fribourg] at 06:28 18 December 2013
488 GARNHAM ET AL.
Materials and Design
The materials and design were based on those of Gygax et al. (2008). For
that study, 36 experimental passages were constructed in English, French, and
German. Each passage comprised two sentences. The first sentence introduced
agroupofpeopleusingarolenameinthepluralform,andthesecond sentence
specified that there were some (but not exclusively) men or some women in the
group (i.e., it provided a partial constraint on the genders of the people in the
group). The participant’s task was to read each passage, presented one sentence
at a time, and to decide, for each sentence pair, whether the second sentence was
asensiblecontinuationofthefirstone.Thedependentvariables were, therefore,
the time to make a judgment of whether the continuation was sensible (only
data for “yes” responses were analyzed) and the proportion of“yesresponses.
An example of a passage is as follows:
1st sentence: The electricians were walking down the street.
2nd sentence: Since sunny weather was forecast, several of the women [men]
weren’t wearing a coat.
In each language, there were 12 stereotypically female role names, 12 stereo-
typically male role names, and 12 neutral role names (chosen from the norms
collected by Gabriel et al., 2008). In French and German, the role names
appeared in the masculine form, which is supposed to be, as a grammatical
rule, interpreted as generic and not as specific. Six different content types
were used for the first sentences. The first sentence mentionedagroupof
people either (a) coming out of a place, (b) waiting somewhere, (c) going into
aplace,(d)beingsomewhere,(e)walking,or(f)goingacrossaplace.For
each content type, there were six different versions—for example, for walking,
walking through the station, and walking across the street. The role names
were randomly assigned to the contents. The second sentencesdifferedfirst,
and most important, in their mention of women or men. Each participant saw
18 continuations about women (6 following sentences with a female-stereotyped
role name, 6 following sentences with a neutral-stereotypedrolename,and
6followingsentenceswithamale-stereotypedrolename)and18aboutmen.
Furthermore, there were three types of continuation content: one based on
different emotions (angry, sad, happy, and joyful), one based on different weather
conditions (sunny, put some sun cream on, cloudy, and need an umbrella),
and one based on different actions (go, have a break, leave, and rest). In all
experimental conditions, the intended response was “yes” (the second sentence
is a sensible continuation of the first). To ensure that the participants read the
passages, 36 filler texts, requiring “no” answers, were constructed. These filler
pairs were similar to the experimental ones (but using different role names), but
Downloaded by [Unviersité de Fribourg] at 06:28 18 December 2013
GENDER REPRESENTATION AND PRONOUNS 489
included a clear semantic or pragmatic incongruity, leadingtoa“noanswer.
The following is an (English) example of such a pair:
(a) The professors were taking a break in the sun.
(b) Due to the bad weather, the majority of the women had an umbrella.
Adding pronouns. Each original first sentence from Gygax et al. (2008)
was extended by adding a pronoun together with some extra (gender neutral)
information. Therefore, all first displays in this experiment introduced a role
name at the beginning and later contained a plural pronoun (they, ils,orsie), as
in the following example:
(a) The neighbors came out of the cafeteria. They went away.
(b) Because of the cloudy weather, one of the women [men] had anumbrella.
(a) Les voisins sortirent de la cafeteria. Ils partirent.
(b) A cause du temps nuageux un[e] des femmes [hommes] avait unpara-
pluie.
(a) Die Nachbarn kamen aus der Cafeteria heraus. Sie gingen weg.
(b) Wegen des bewölkten Wetters, hatte eine[r] der Frauen [Männer] einen
Regenschirm.
Control Task for a Possible Alternative Interpretation of the
Passages
The determiners in the second sentences were meant to be inclusive (i.e., “some
of the women” means that there could be men in this group as well). However,
they could be interpreted as exclusive, hence biasing responses. For example,
when reading the following German passage, there are two reasons why the
second sentence (b) might be judged as not a sensible continuation of the first
one:
(a) Die Statistiker passierten die Strasse.
‘The statisticians were walking into the street’.
(b) Weg en d e r Hi tz e t ran k e in e der Frau en Wa ss e r.
‘Because of the heat, one of the women was drinking water’.
As Statistiker is stereotypically male, participants could think that it isnot
possible that a woman is a statistician. However, participants may also think that
“one of the women” means that there are only women in the group (i.e., they
might take “one of ” to induce an exclusive interpretation of the composition
of the group). In this case, there is a grammatical mismatch between the word
“Statistiker” (masculine grammatical gender) and (the correct way of referring
Downloaded by [Unviersité de Fribourg] at 06:28 18 December 2013
490 GARNHAM ET AL.
to) a group composed only of women, which should have been die Statistik-
erinnen ‘the female statisticians’. What a participant takes to be a grammatical
mismatch might be interpreted as a stereotype mismatch.
To i nve s ti g at e w het her peo pl e m ake t hi s seco nd in ter p ret ation, we introduced
afurthershorttask.Thistaskwasadministereddirectlyafter the main exper-
iment, and contained 12 passages (6 experimental and 6 fillers). All passages
were similar to those in the main experiment. A first display (comprised of
2shortsentences)wasfollowedbyasecondone,andtheparticipant’s task was
to judge if the sentence in the second display was a sensible continuation of the
first. The first display (a) mentioned that some people were doing something, and
among them there was one woman or one man. A second sentence (b)referred
to some men or women (the opposite of the first sentence), whosenumberis
given by one of the six determiners (in this case, “some”):
(a) The people came out of the room. One woman was wearing a raincoat.
(b) Because of the bad weather, one of the men had an umbrella.
The expected answer was “yes” in all experimental passages. However, if the
participants thought that “one of the men” meant that there were only men in the
group of people, they would give a negative answer. To make sure participants
attended to the task, six filler passages were added with “no” as the correct
answer. There were two types of filler passages (see the following examples).
Some stated in the first sentence that they were people of only one gender (c),
and then the other gender was mentioned in the second sentence(d).Otherfiller
passages were semantically or pragmatically incongruent (e&f):
(c) The group of men went into the building. One man looked at the mail-
boxes.
(d) After so little time, a few of the women seemed to want to go on.
(e) The people were at the airport. One man seemed happy.
(f) One could see that several of the women were swimming.
Before analyzing the data from the main experiment, we examined whether
all determiners were interpreted as inclusive. We decided that for a determiner to
be considered inclusive, it should produce more than 50% of the positive answers
in the control task. On this basis, as shown in Table 2, all determiners in all
languages were interpreted inclusively (Min D59%, Max D94%). Although
all analyses for the main experiment were run including all responses, we also
reran the analyses considering, for each participant, only those determiners that
were considered by them as inclusive in the control task (i.e., above 50% for
each participant). In addition, we ran analyses in which the determiners were
included as an additional variable. We only present the first set of analyses (all
Downloaded by [Unviersité de Fribourg] at 06:28 18 December 2013
GENDER REPRESENTATION AND PRONOUNS 491
TABLE 2
Percentage of Positive Answers Accompanying Sentences Containing the
Different Determiners in All Three Languages
English % French % German %
Afewofthe 94 Quelques 94 Einige 94
The majority of the 82 La majorité des 68 Die Mehrheit der 66
Most of the 61 La plupart des 61 Die meisten 59
One of the 91 Une/un des 94 Eine/einer der 91
Several of the 70 Plusieurs 65 Mehrere 63
Some of the 76 Une partie des 68 Ein Teil der 66
M79 75 74
responses and without determiners as a variable), as the results of the second
and third did not differ from those of the first.
Procedure for Main Experiment
The participants were individually tested in a small, quiet room. Their task was to
read each passage, presented in two parts, and to decide whether the second part
was a “sensible” continuation of the first one. In French and German, we used
the terms une continuation possible and eine mögliche Fortsetzung,respectively,
which we judged to be semantically closest to the English wordsensible. The
participants in all languages were asked to make a quick decision based on their
first impression, and not on prolonged reflection. A prompt (i.e., **Ready?**,
**Prêt?**, and **Bereit?**) appeared on the screen before each passage. The
participants pressed the “yes” button to make the first display appear, and then
pressed the “yes” button again to make the second display (target sentence)
appear. They then had to make a prompt decision by pressing either the “yes”
button (i.e., “I think it’s a sensible continuation”) or the “no” button (i.e., “I
don’t think it’s a sensible continuation”). Participants were asked to keep the
index finger of their dominant hand on the “yes” button and the index finger of
their non-dominant hand on the “no” button.
RESULTS
We pred i cted t hat in En gli sh the r e woul d be a s t ereo t yp e m atch–mismatch
effect (Stereotype !Continuation interaction), which would be the same in this
experiment as in Gygax et al. (2008). In French and German, we predicted a main
effect of continuation (male continuations more easily processed) with a possible
Downloaded by [Unviersité de Fribourg] at 06:28 18 December 2013
492 GARNHAM ET AL.
modulation in each language (Experiment !Continuation interaction). These
effects do not neatly map onto interaction effects in an overall analysis with
language as a factor; and, indeed, none of the relevant higher-order interactions
with language were significant in such analyses. However, because we made
specific predictions for each language, we present separate analyses for the
three languages that specifically test the predictions we made. The sample sizes
in Gygax et al. are directly comparable to those in this study,andwere35,36,
and 36 for English, French, and German, respectively.
Proportion of Positive Judgments
The mean proportions of positive judgments in each of the three languages
are shown in Table 3, along with the means from Gygax et al. (2008). To
compare the results of this experiment and those of Gygax et al., we conducted
both by-participants (F1) and by-items analyses (F2; see Clark, 1973). In the
former (F1), mixed-design analyses of variance (ANOVAs) were conducted with
stereotype (male vs. female vs. neutral) and continuation (men vs. women)
as within-subjects variables and experiment (Gygax et al. vs. this experiment)
as a between-subject variable. In the latter (F2), mixed-design ANOVAs were
conducted with experiment (Gygax et al. vs. this experiment) and continuation
TABLE 3
Mean Proportions of Positive Judgments (and Standard Deviations) in
English, French, and German as a Function of Stereotypes and Continuations in
This Experiment and Gygax, Gabriel, Sarrasin, Oakhill, and Garnham (2008)
Continuation: Contained the Word “Men” or “Women
This Experiment Gygax et al.
Men Women Men Women
Language Stereotype M SD M SD M SD M SD
English Female .65 .31 .81 .22 .65 .32 .88 .20
Male .81 .21 .65 .30 .85 .16 .66 .26
Neutral .75 .23 .69 .31 .81 .21 .81 .28
French Female .76 .29 .62 .29 .77 .28 .59 .32
Male .83 .20 .54 .30 .83 .23 .58 .29
Neutral .76 .25 .66 .25 .73 .34 .56 .34
German Female .68 .31 .57 .30 .65 .33 .40 .28
Male .79 .19 .46 .30 .69 .29 .35 .33
Neutral .80 .26 .64 .29 .72 .28 .45 .32
Downloaded by [Unviersité de Fribourg] at 06:28 18 December 2013
GENDER REPRESENTATION AND PRONOUNS 493
(men vs. women) as within-items variables (i.e., the same role names were
presented in each experiment, and each role name was followedbymaleand
female continuations) and stereotype (male vs. female vs. neutral) as a between-
item variable.
English data. The analysis revealed a main effect of experiment when
considering items as a random factor, F1(1, 69) D1.51, ns and F2(1, 33) D
14.55, p<.01: There were fewer positive answers in this experiment (73%) than
in Gygax et al. (2008; i.e., 78%). Most important, there was nomaineffectof
continuation, F1(1, 69) <1andF2(1, 33) <1; nor of stereotype, F1(2, 138) <1
and F2(2, 33) <1. However, as expected, there was a Stereotype !Continuation
effect, F1(2, 138) D31.28, p<.001 and F2(2, 33) D28.78, p<.001; and
no Stereotype !Continuation !Experiment effect, F1(2, 138) <1andF2(2,
33) <1. There were more positive judgments for continuations thatmatchedthe
stereotype than for those that did not; for neutral role names, both continuations
were equally accepted. The results found in this experiment replicated those
found in Gygax et al., suggesting that, in English, the mentalrepresentationof
gender when reading role names is solely based on the stereotypicality of those
role names.
French data. The analysis revealed no main effect of experiment, F1(1,
67) <1andF2(1, 33) D1.33, ns;butamaineffectofcontinuation,F1(1, 67) D
79.84, p<.001 and F2(1, 33) D48.78, p<.001, supporting an overall male
bias. There was no Experiment !Continuation effect, F1(1, 67) <1andF2(1,
33) <1. The presence of a masculine pronoun in addition to the role name in the
grammatically masculine form in the priming sentences did not increase the male
bias found earlier, although our prediction was only that thebiasshouldatleast
be maintained. Thus, for French, the results from this study and those from our
previous study were very similar. There was also a Stereotype!Continuation
effect when considering participants as a random factor, F1(2, 134) D5.84,
p<.05 and F2(2, 33) D1.89, ns: The difference between male and female
continuations was slightly bigger in the male stereotype condition (27%) than
in the female stereotype condition (16%) and the neutral stereotype condition
(14%).
German data. The analysis revealed a main effect of experiment, F1(1,
70) D7.30, p<.05 and F2(1, 33) D27.86, p<.001: There were more positive
answers in this experiment (66%) than in Gygax et al. (2008; i.e., 54%). There
was a strong main effect of continuation, F1(1, 70) D74.09, p<.001 and F2(1,
33) D143.04, p<.001: There were more positive answers to men continuations
than to women continuations. Most crucially, this main effect was qualified by
an Experiment !Continuation interaction effect, but only when considering
Downloaded by [Unviersité de Fribourg] at 06:28 18 December 2013
494 GARNHAM ET AL.
items as a random factor, F1(1, 70) D2.50, ns and F2(1, 33) D8.88, p<.01,
revealing an attenuated male bias in this experiment compared to Gygax et al.
(i.e., a difference in positive judgments between men vs. women continuations:
this experiment D20% and Gygax et al. D29%).2As we expected, the presence
of a pronoun morphologically identical to the feminine singular seems to have
facilitated positive answers to continuation sentences about women, at least when
considering items as random factor.
There was also a main effect of stereotype, F1(2, 140) D4.82, p<.05 and
F2(2, 33) D3.62, p<.05, showing more positive answers to the neutral role
names (65%) than to the male (57%) and female (57%) stereotyped role names
(least significant difference [LSD] with p<.05), as well as a Stereotype !
Continuation effect, F1(2, 140) D4.24, p<.05 and F2(1, 33) D5.96, p<
.01. Po s t ho c analyses (LSD) showed that the difference in positive judgments
between men and women continuations was higher for the male stereotyped
role names (34%) than for the female (18%) and neutral (21%) stereotyped role
names (p<.05). On a descriptive level, when comparing the two experiments
(see Figure 1, solid lines), the attenuation of the masculinebiasseemedmore
pronounced in the female (decrease of 15%) and neutral stereotyped conditions
(decrease of 11%) than in the male stereotyped condition (decrease of 1%).
Judgment Times
Only response times for positive judgments were analyzed. The proportion of
positive responses was quite low in some conditions, which led to an imbalanced
dataset. To accommodate this problem the data were analyzed by fitting linear
mixed-effects models (Statistical Package for the Social Sciences 18.0: SPSS,
Inc., Chicago, IL), including both participants and items asrandomfactors
(Brysbaert, 2007). As in Gygax et al. (2008), judgment times that were 2.5 SD
or more above each participant’s mean were replaced by the 2.5SD cutoff (1%
of French, 2% of German, and 2% of English times were affected). All means
for judgment times in both this experiment and in Gygax et al. are shown
in Table 4. Separate models were estimated for each language.Experimental
factors (stereotype, continuation, experiment, and their interactions) were treated
2This result was confirmed in an ipsative analysis in which the number of positive answers
given to men in each stereotype condition and for each participant was divided by the total
number of positive answers for the specific stereotype condition, and similarly for the number
of positive answers to women continuations. Regarding continuations, the resulting scores for men
and women always sum to 1 for each person in each stereotype condition. A 2 (Experiment) !3
(Stereotype) mixed-design analysis of variance was run on the proportions of positive judgments for
men continuations. In this ipsative data analysis, main effects of experiment and stereotype mirror
Experiment !Continuation and Stereotype !Continuation interactions in the main analysis. In this
analysis, the main effect of experiment was significant, F(1, 69) D4.77, p<.05.
Downloaded by [Unviersité de Fribourg] at 06:28 18 December 2013
GENDER REPRESENTATION AND PRONOUNS 495
FIGURE 1 Differences in proportions of positive judgments for men versus women
continuations as a function of languages and stereotypes in this experiment and Gygax,
Gabriel, Sarrasin, Oakhill, and Garnham (2008).
TABLE 4
Judgment Times of Positive Answers (and Standard Deviations) in
English, French, and German as a Function of Stereotypes and Continuations in
This Experiment and Gygax, Gabriel, Sarrasin, Oakhill, and Garnham (2008)
Continuation: Contained the Word “Men” or “Women
This Experiment Gygax et al.
Men Women Men Women
Language Stereotype M SD M SD M SD M SD
English Female 2,851 1,061 2,862 1,295 2,749 1,004 2,913 1,212
Male 2,661 1,112 2,912 857 2,810 934 2,954 1,164
Neutral 2,824 1,140 2,830 902 2,885 1,476 2,912 961
French Female 2,923 879 3,201 1,278 3,665 1,626 3,875 1,486
Male 2,941 824 3,137 1,667 3,523 1,380 3,866 1,791
Neutral 3,137 1,804 3,125 1,131 3,701 1,101 3,873 1,355
German Female 2,888 923 2,850 1,188 3,088 1,336 3,512 1,657
Male 2,663 1,086 2,851 930 3,016 1,514 3,374 1,466
Neutral 2,688 1,023 2,872 1,202 3,018 1,336 3,559 1,676
Downloaded by [Unviersité de Fribourg] at 06:28 18 December 2013
496 GARNHAM ET AL.
as fixed effects; participants and role names were treated as random effects.
Participants were nested under experiment,androlenameswerenestedunder
stereotype. By means of chi-square difference tests, we explored the change
in model fit when further adding random slopes. In the following sections, we
report just those random effects that improved the fit of the model. In no case
did adding these effects significantly change the conclusions from the fixed part
of the model.
English data. Adding the Stereotype !Participant interaction as a random
effect (i.e., participants differ in their susceptibility to the stereotype information)
significantly improved the fit of the model (!"2D63, !df D1, p<.001). The
largest, but only marginally significant, effect was the Stereotype !Continuation
effect, F(2, 1,742.80) D2.80, pD.06 (all other ps>.20). On a descriptive level,
the Stereotype !Continuation pattern supports the findings from the proportion
judgments: The difference in positive judgment times between men and women
continuations was higher for male stereotyped role names ("277 ms) and for
female stereotyped role names (C129 ms) than for neutral role names ("75 ms).
Thus, for English, the results of this study and our previous one were very
similar.
French data. Adding the Stereotype !Participant interaction as a random
effect (i.e., participants differ in their susceptibility to the stereotype information)
significantly improved the fit of the model (!"2D15, !df D1, p<.001). The
analysis revealed a main effect of experiment, F(1, 65) D5.40, pD.02, showing
that there were faster, positive responses in this experiment (3,049 ms) than in
Gygax et al. (2008; i.e., 3,657 ms), as well as a main effect of continuation,
F(1, 1,554.20) D8.00, p<.01—positive responses being faster for sentences
containing men (3,263 ms) than sentences containing women (3,444 ms). As in
the analysis of the proportion of positive responses, there was no Experiment !
Continuation effect, F(1, 1,561.90) D1.10, pD.34 (all other Fs<1). The bias
reported by Gygax et al. was maintained, but not enhanced. Theonlydifference
between the two studies was the faster mean reaction time in this study, compared
with the previous one.
German data. Adding the Stereotype !Continuation !Participant inter-
action as a random effect (i.e., participants differ in theirsusceptibilitytothe
stereotype-continuation match/mismatch) significantly improved the fit of the
model (!"2D35, !df D1, p<.001). The analysis revealed a main effect
of continuation, F(1, 229.40) D15.60, p<.001—positive responses being
faster for sentences containing men (2,882 ms) than sentences containing women
(3,123 ms). There was also a main effect of experiment, F(1, 69.27) D5.40,
pD.02, showing that there were faster positive responses in this experiment
Downloaded by [Unviersité de Fribourg] at 06:28 18 December 2013
GENDER REPRESENTATION AND PRONOUNS 497
(2,790 ms) than in Gygax et al. (2008; i.e., 3,205 ms). Most crucially, as expected
and supporting the analysis of the proportion of positive responses, there was an
Experiment !Continuation effect, F(1, 753.20) D5.70, pD.02, suggesting an
attenuated male bias in this experiment compared to Gygax et al. The difference
in positive judgment times between men and women continuations was higher
in Gygax et al. (451 ms) than in this experiment (120 ms). The presence of
apronounidenticaltothefemininesingularseemstohavefacilitated positive
response times to target sentences containing women (all other Fs<1).
DISCUSSION
The objective of this research was to examine the influence, ifany,ofthe
different grammatical systems in French and German on the grammatical bias
in gender interpretation found in Gygax et al. (2008). The central idea was to add
referential pronouns (they, ils,andsie)thatcarrieddifferentgrammaticalgender
cues in the languages under scrutiny. Our study then investigated whether the use
in French of the pronoun ils would maintain or reinforce the masculine bias and
if the use of sie in German would attenuate it. In English, we did not expect any
changes, as the pronoun they is not gender marked. An alternative possibility was
that the basic grammatical information on the pronouns is used by the anaphor
resolution process, but that the more subtle aspects of the morphological marking
(i.e., the relation between the form actually used and other forms in the language)
should not affect the already-established representation of the pronoun’s referent.
In line with our expectations in English, the proportions of positive judgments
and the positive judgment times revealed that the gender representation was
biased by stereotyped information (or lack of it, in the case of the neutral items,
so that they readily maps onto the representation of, say, “singers, andboth
“men” and “women” in the second sentence are seen as equally consistent with
that representation), as in Gygax et al. (2008). Also in line with our expectations
in German, the potentially feminine form of the plural pronoun sie significantly
weakened the overall male bias reported by Gygax et al. for both the proportion
of positive judgments and the judgment times, as well as for all three types
of role nouns including, most important, neutral ones. This finding is in line
with Rothermund’s (1998) results and, thus, provides the first corroboration
of his post hoc explanation of those results: The German plural pronoun sie,
which is morphologically identical to the feminine singularpronoun,hasfemale
associations that work against the male associations evokedbythemasculine
role names. In French, the addition of pronouns had no effect;however,because
those cues provided the same information as the GM, we could not definitively
predict an enhancement effect, as opposed to maintenance of the strength of
bias in the gender representation. By adding further grammatical cues, we
Downloaded by [Unviersité de Fribourg] at 06:28 18 December 2013
498 GARNHAM ET AL.
succeeded in attenuating (in German), but not in further amplifying (in French),
the masculine bias brought about by employing the masculine (supposedly
generic) forms of the role names. In short, combining grammatical cues that
match in their gender marking (as in French; i.e., a masculinenounanda
masculine pronoun) does not seem to have an additive effect, whereas combining
grammatical cues that do not match (as in German; i.e., a masculine noun and
apluralpronounmorphologicallyidenticaltoafemininesingular one) seems to
distract readers from forming a specifically male gender representation. More
generally, the question of whether the gender-based part of the representation
of a person, introduced by a referential noun phrase and comprising (possibly
morphologically gender marked) determiner and role name, can be modulated
by gender information in a later pronoun can be answered in thepositive,at
least in the German case. As noted in the introduction of this article, the result,
although compatible with the overall mental models framework and its strictures
about the nature of representations of content, shows that other influences are at
work in the language processing system. More specifically, the relation between
the German plural sie and the German feminine singular sie is independent of
what is referred to by a particular use of the German plural sie (all women,
all men, or a mix of men and women). This result indicates another case,
like that of stereotype information where information that is not necessary for
comprehension and, indeed, may be detrimental to comprehension in certain
cases, is nevertheless activated during comprehension.
In relation to the use of stereotype information in German andFrench,in
German we found a Stereotype !Continuation effect for the proportions of
positive judgments. Whereas the masculine bias (more positive answers for men
continuations than for women continuations) was of a similarstrengthforfemale
stereotypical and neutral role names, it was stronger for male stereotyped role
names. Although we expected dissimilarities between the German and French
data due to the use of different grammatical cues, the patternofresultsdisplayed
in Figure 1 (and also Table 3) suggests a similar effect for theFrenchsample
(see Figure 1, dotted lines) in this experiment and the corresponding sample in
Gygax et al. (2008). When analyzing the proportion of positive judgments from
the French sample of this experiment on its own, a significant Continuation !
Stereotype effect—F1(2, 66) D4.87, p<.05 and F2(2, 33) D3.25, p<.06
emerged, corroborating the notion of a numerically similar trend to the German
sample. Together, these results may indicate an influence of (male) stereotype
information on gender representation both in German and (less pronounced) in
French.
Because we presented whole sentences and did not examine, in detail, times
for reading the first part of the passages, our results do not directly bear on the
time course of gender processing. Nevertheless, previous research suggests that,
with the kinds of passages and procedure we used, processing and encoding of
Downloaded by [Unviersité de Fribourg] at 06:28 18 December 2013
GENDER REPRESENTATION AND PRONOUNS 499
stereotype information and morphological information on nouns and pronouns
is more or less immediate (Carreiras et al., 1996; Duffy & Keir, 2004).
To con clu de, w e b eli eve tha t t hi s exp eri ment ha s fu rt h ere d o ur understanding
of the interaction between grammatical cues and stereotypical information when
constructing a representation of gender during reading. In non-gender-marked
languages, such as English, readers based their representations on stereotype in-
formation. In gender-marked languages, when pronouns are added, subtle aspects
of their morphology, such as the identity of the plural and thefemininesingular
form (sie)inGerman,modifythegenderrepresentationoftheantecedents of
those pronouns.
ACKNOWLEDGMENT
This research was supported by Grant No. 100014-118301 from the Swiss
National Foundation.
REFERENCES
Académie Française. (2002). Féminisation des noms de métiers, fonctions, grades et titres [Gram-
matical feminization of professions, function, grades, andtitles].Paris,France:Author.Retrieved
from http://www.academie-francaise.fr/actualites/feminisation.asp
Baudino, C. (2001). Po l i tiq u e d e la l an g u e e t d iffé r e n c e s e x u e l le : L a p o l i tis a tion du genre des noms
de métiers [Language politics and sexual difference: The politicization of job titles’ gender]. Paris,
France: L’Harmattan.
Brauer, M., & Landry, M. (2008). Un ministre peut-il tomber enceinte? L’impact du générique
masculin sur les représentations mentales [Can a secretary of state become pregnant? The impact
of the generic masculine on mental representations]. L’An e Psycholo gique, 108 , 243–272.
Braun, F., Gottburgsen, A., Sczesny, S., & Stahlberg, D. (1998). Können Geophysiker Frauen sein?
Generische Personenbezeichnungen im Deutschen [Can geophysicians be women? Generic terms
in German]. Zeitschrift für Germanistische Linguistik, 26, 265–283.
Braun, F., Sczesny, S., & Stahlberg, D. (2005). Cognitive effects of masculine generics in German:
An overview of empirical findings. Communications, 30, 1–21.
Brysbaert, M. (2007). “The language-as-fixed-effect fallacy”: Some simple SPSS solutions to a
complex problem (Version 2.0). Egham, England: Royal Holloway, University of London.
Carreiras, M., Garnham, A., Oakhill, J., & Cain, K. (1996). The use of stereotypical gender
information in constructing a mental model: Evidence from English and Spanish. Quarterly
Journal of Experimental Psychology Section A—Human Experimental Psychology, 49, 639–663.
Chatard, A., Guimond, S., & Martinot, D. (2005). Impact de la féminisation lexicale des professions
sur l’auto-efficacité des élèves: Une remise en cause de l’universalisme m asculin? [Occupational
self-efficacy as a function of grammatical gender in French].L’Ann ée Psych ologique, 105, 249–
272.
Clark, H. H. (1973). The language-as-fixed-effect fallacy: Acritiqueoflanguagestatisticsin
psychological research. Journal of Verbal Learning and Verbal Behavior, 12, 335–359.
Duden: Die Grammatik [The German Grammar, 8th ed.]. (2009). Mannheim, Germany: Dudenverlag.
Downloaded by [Unviersité de Fribourg] at 06:28 18 December 2013
500 GARNHAM ET AL.
Duffy, S. A., & Keir, J. A. (2004). Violating stereotypes: Eyemovementsandcomprehension
processes when text conflicts with world knowledge. Memory & Cognition, 32, 551–559.
Gabriel, U. (2008). Language policies and in-group favouritism: The malleability of the interpretation
of generically intended masculine forms. Social Psychology, 39, 103–107.
Gabriel, U., & Gygax, P. (2008). Can language amendments change gender representation? The case
of Norway. Scandinavian Journal of Psychology, 49, 451–457.
Gabriel, U., Gygax, P., Sarrasin, O., Garnham, A., & Oakhill,J.(2008).Au-pairsarerarelymale:
Norms on the gender perception of role names across English, French and German. Behavior
Research Methods, 40, 206–212.
Garnham, A., & Oakhill, J. (1996). The mental models theory oflanguagecomprehension.InB.K.
Britton & A. C. Graesser (Eds.), Models of understanding text (pp. 313–339). Hillsdale, NJ:
Lawrence Erlbaum Associates.
Garnham, A., Oakhill, J., & Reynolds, D. (2002). Are inferences from stereotyped role names to
characters’ gender made elaboratively? Memory & Cognition, 30, 439–446.
Gygax, P., & Gabriel, U. (2008). Can a group of musicians be composed of women? Generic
interpretation of French masculine role names in absence andpresenceoffeminineforms.Swiss
Journal of Psychology, 67, 143–151.
Gygax, P., Gabriel, U., Sarrasin, O., Oakhill, J., & Garnham,A. (2008).Genericallyintended,but
specifically interpreted: When beauticians, musicians and mechanics are all men. Language and
Cognitive Processes, 23, 464–485.
Irmen, L., & Roßberg, N. (2004). Gender markedness of language: The impact of grammatical and
nonlinguistic information on the mental representation of person information. Journal of Language
and Social Psychology, 23, 272–307.
Kennison, S. M., & Trofe, J. L. (2003). Comprehending pronouns: A role for word-specific gender
stereotype information. Journal of Psycholinguistic Research, 32, 355–378.
Oakhill, J., Garnham, A., & Reynolds, D. (2005). Immediate activation of stereotypical gender
information. Memory & Cognition, 33, 972–983.
Rothermund, K. (1998). Automatische geschlechtsspezifische assoziationen beim lesen von texten
mit geschlechtseindeutigen und generisch maskulinen text-subjekten [Automatic gender-specific
associations to texts containing gender-specific and masculine generic text subjects]. Sprache &
Kognitio n, 17, 183–198.
Rothmund, J., & Scheele, B. (2004). Personenbezeichnungsmodelle auf dem prüfstand: Lösung-
smöglichkeiten für das genus-sexus-problem auf textebene [Putting gender-neutral reference terms
to the test: Constructive solutions to the problem of grammatical versus referential gender on the
textual level]. Zeitschrift Für Psychologie, 21, 40–54.
Stahlberg, D., Braun, F., Irmen, L., & Sczesny, S. (2007). Representation of the sexes in language.
In K. Fiedler (Ed.), Social communication (pp. 163–187). New York, NY: Psychology Press.
Sturt, P. (2003). The time-course of the application of binding constraints in reference resolution.
Journal of Memory & Language, 48, 542–562.
Tan e n h a u s , M . K., & C a r l s o n , G . N. ( 19 9 0 ) . C o mp r eh e n s i o n of d eep and surface verb phrase
anaphors. Language and Cognitive Processes, 5, 257–280.
Zwaan, R., & Radvansky, G. (1998). Situation models in language comprehension and memory.
Psychological Bulletin, 123, 162–185.
Downloaded by [Unviersité de Fribourg] at 06:28 18 December 2013
... Several studies have shown that girls and/or women are more likely to be (mentally) associated with the given subject when gender-equitable language is used in the corresponding texts compared to those texts without gender-equitable language (e.g. Garnham et al. 2012). ...
... The generic masculine is used for different purposes, which partly leads to confusion, since, for example, in one picture with only male persons and in a directly following one with male and female persons, the term chicos is used (3). The generic masculine in the plural is also found in numerous other formulations (4,18,19,20), but is only used JOURNAL OF LANGUAGE TEACHING AND RESEARCH 215 consequently and coherently for this purpose in the textbook ¡ Vamos! ¡ Adelante! 2, which is why there is no form of gender specification or abstraction in the tasks. ...
Article
In the following article, gender-related aspects in ELE textbooks will be examined from a broader thematic point of view, as well as detailed linguistic perspective. As a starting point, the state of research on gender-related studies in foreign language teaching pedagogy and research in Germany will be outlined, followed by a discussion on the importance of gender awareness promoting textbooks. After a general introduction and a rationale for the relevance of the study, central findings of studies examining gender representation in ELE textbooks will be discussed. This is followed by an analysis, in which specifics of textbook analysis and sampling criteria are explained. Furthermore, an analysis model is developed, based on the latest findings of gender-related textbook research. After an excursus on language-specific possibilities of gender sensitivity in Spanish, the results of the analysis are presented and discussed. Finally, implications of the findings for teachers will be deduced and recommendations regarding what can be done to promote and deepen gender awareness of learners will be given.
... • however, previous research has cast doubt on the gender-neutral use of masculine generics • most (if not all) behavioural studies on the subject find one overall result → masculine generics are not gender-neutral but show a clear bias towards the explicit masculine reading (e.g. Demarmels, 2017;Garnham et al., 2012;Gygax et al., 2008;Irmen & Kurovskaja, 2010;Irmen & Linner, 2005;Koch, 2021;Misersky et al., 2019;Stahlberg & Sczesny, 2001;Trutkowski, 2018) • even though a masculine generic may be used by a speaker with the intention of considering all genders… • …this intention is not fully translated by the receiver's comprehension system ...
... • the explicit feminine is more similar to the explicit masculine than to masculine generics • masculine generics show a clear bias towards the masculine reading, producing a 'male bias' in the language system itself • thus, our findings confirm the bias found in previous behavioural studies (e.g. Demarmels, 2017;Garnham et al., 2012;Gygax et al., 2008;Irmen & Kurovskaja, 2010;Irmen & Linner, 2005;Koch, 2021;Misersky et al., 2019;Stahlberg & Sczesny, 2001;Trutkowski, 2018) • future research will show ...
Conference Paper
Full-text available
Research of the last two decades has repeatedly shown that masculine generics in German exhibit a masculine bias (e.g. Gabriel et al., 2008; Gygax et al., 2008; Irmen & Kurovskaja, 2010; Koch, 2021; Misersky et al., 2019; Stahlberg & Sczesny, 2001). That is, grammatically masculine role-nouns (e.g. Lehrer, ‘teacher’ / ‘teachers’) can refer to men and women, but may favour an interpretation in which only men are considered as potential referent (Misersky et al., 2019). While previous studies are consistent in finding such a male bias for masculine generics, very few of them offer a theoretical account on the underlying nature of this bias (e.g. Irmen & Linner, 2005). Up to this point, no attempt has been made to find out whether there are connections between the male bias and the representations of masculine generics in the mental lexicon. This is what the present paper offers. Using the general ideas of distributional semantics (e.g. Harris, 1954) as a framework, the following questions are explored: Are masculine generics generic or do they show a bias? If a bias is found, how can we explain it in terms of underlying representations in the mental lexicon? To answer this question, the following method was employed. An 830,000 sentence (1.7 million words) corpus of contemporary German was created using the Leipzig Corpora Collection (Goldhahn et al., 2012). The corpus included 120 target word pairs which were based on the set of words used in Gygax et al. (2008). Target words were grammatically masculine role nouns and their grammatically feminine counterparts. All target word occurrences were manually checked for their usage, i.e. whether they were generically or specifically intended, and annotated accordingly. The corpus was then used to train semantic vectors based on the Rescorla-Wagner equation (Wagner & Rescorla, 1972) as implemented by naïve discriminative learning (e.g. Baayen & Ramscar, 2015). As a general result, the semantic vectors of masculine generics show a bias towards the grammatical masculine, i.e. the masculine specific. Further cosine similarity analyses show that the generic masculine is overall semantically more similar to words denoting males, e.g. Mann, ‘man’. A rank order correlation analysis corroborates this finding. Our results indicate that the male bias of masculine generics in German is due to the underlying representations of masculine role-nouns in the mental lexicon. That is, even though they may be intended to be generic, their resonance with the lexicon, i.e. more specifically with masculine specifics, leads to an overall biased association with male referents.
... The finding that gender stereotypes informed participants' responses in the Swiss gendered version is in keeping with research suggesting that grammatical gender and stereotypes do interact (e.g., Irmen, 2007;Irmen & Roßberg, 2004;Vervecken et al., 2015), while the finding that it did not modulate participants' responses in the Canadian gendered version is keeping with other research that found the male bias to completely override stereotype effects (e.g., Garnham et al., 2012;Sato et al., 2016). We believe that the paradigm used in the present study (i.e., using first names) may potentially increase French speakers' sensitivity to gender information more generally, and in potentially different manners depending on specific cultural context. ...
Article
Full-text available
The extent to which gender neutral and gendered nouns impact differently upon native French speakers' gender representations was examined through a yes-no forced choice task. Swiss (Experiment 1) and Québec (Experiment 2) French-speaking participants were presented with word pairs composed of a gendered first name (e.g., Thomas) and a role (e.g., doctor), and tasked to indicate whether they believed that [first name] could be one of the [role]. Roles varied according to gender stereotypicality (feminine, masculine, non-stereotyped), and were either in a plural masculine (interpretable as generic) or gender neutral (epicenes and group nouns) form. The results indicated that the use of gender neutral forms of roles avoided a strong male bias found for the masculine forms, and that both gender neutral and masculine forms used equal cognitive resources. Further, stereotype effects associated with both gender-neutral and grammatically masculine forms were quite small (<1%). These results were highly reliable across both Swiss French and Québec speakers. Our study suggests that gender neutral forms are strong alternatives to the use of the masculine form as default value.
... Psycholinguistic research on the interpretation of gender marking in French is much more recent (see Gygax et al., 2013, for an overview). However, confirming literary and grammatical studies going back to Beauvoir (1949) and Yaguello (1979), Houdebine (1987), and Michard (1996), a large and growing body of work on this language has shown that French behaves similarly to English: It is almost impossible to refer both men and women in an equal way by using a masculine marked expression (Brauer & Landry, 2008;Chatard et al., 2005;Gabriel et al., 2008;Garnham et al., 2012;Gygax et al., 2008Gygax et al., , 2012Gygax et al., , 2019, among many others). Grammatically masculine noun phrases have an interpretative bias in favor of men that goes above and beyond the particular stereotypes associated with the noun. ...
Article
Full-text available
Écriture inclusive (EI) has long been the topic of public debates in France. These debates have become more intense in recent years, often focusing on the higher education system and culminating in the formulation of three separate laws banning it for public administration. In this paper, we investigate the foundations of these conflicts through a large quantitative corpus study of the (non)use of EI in Parisian undergraduate brochures. Our results suggest that Parisian university professors use EI not only to ensure gender neutral reference but also as a tool to construct their political identities. We show that both the use of EI and its particular forms are conditioned by how brochure writers position themselves on non gender-related-related issues within the French university's political landscape, which explains how conflicts surrounding a linguistic practice have become understood as conflicts about larger issues in French society. Our paper thus provides new information to be taken into account in the formulation and promotion of nonsexist language policies and sheds light on how feminist linguistic activism and its opposition are deeply intertwined with other kinds of social activism in present-day France.
... Future research could conduct a deeper analysis of text structure-for example, by examining storylines-to disentangle the activation of gender stereotypes through written material (cf. studies regarding grammatical gender representations: Garnham et al. 2012). ...
Article
Full-text available
Societies have socially shared assumptions about what constitutes typically male or female attributes. Language can contribute to gender inequality by transmitting gender stereotypes. This study examines whether gender-stereotypical connotations in stimulus texts within a reading competence test might serve as a nuisance factor distorting reading competence measurements. In addition to a general factor for reading competence, we expected gender-stereotypical texts to give rise to gender-specific factors regarding the text content. The research was based on a sample of 813 adults from a pilot study of the German National Educational Panel Study (NEPS). A bifactor model confirmed a general factor for reading competence. However, the two gender-specific factors were not found; consequently, no substantial gender differences in reading competence for gender-stereotypical text content were observed. These findings indicate that there is no substantial impact of gender-stereotypical text connotations on the measurement of women’s and men’s reading competence.
... Consequently, GFL-form recommendations have recently been published for use in higher education (Guerrero Salazar, 2013). As a result of those actions, the data on the use of GFL in universities and its influence on students and lectures shows a positive trend in some societies and for some languages (Garnham et al., 2012;Sarrasin et al., 2012). From a longitudinal perspective, an increase in the use of GFL has been reported (Nissen, 2013), and there are positive results on the acceptance of some GFL forms among university students, such as the use of @ to refer to both genders or, in the Spanish case, a preference for epicene nouns (Bengoechea & Simón, 2014). ...
Article
The article analyses the use of GFL in a corpus of 187 academic texts created by pre-service teachers. It reveals how participants reached a balance between the recommendations of GFL guides and the standard normative grammar included in the school curriculum. The study shows that although future teachers are aware of GFL and sensitive to discrimination, they used a combination of the generic masculine with GFL recommendations when deemed necessary. This has great pedagogical potential for the Spanish classroom, together with the potential to transform language and society.
Book
Full-text available
Der Band vereint neueste, empirisch abgesicherte Forschung zu den Zusammenhängen zwischen Genus, Sexus und der sozialen Kategorie Gender. Reflektiert werden sprachliche Erscheinungen auf der Wortebene, im Bereich der Syntax und Textkohärenz und soziopragmatische sowie diskurstheoretische Fragen. Die wortzentrierten Beiträge umfassen Untersuchungen zu genderstereotypen Bedeutungsangaben in Wörterbüchern, zur in-Movierung bei Anglizismen und zur Reichweite des Genus-Sexus-Prinzips bei Tierbezeichnungen. Syntaktische Fragestellungen zielen auf anaphorische Wiederaufnahmephänomene und deren innersprachliche und pragmatische Bedingungen. Dies betrifft hybride Nomina und Epikoina (das Kind > es/?er; das Mädchen > es/sie) ebenso wie Koreferenzphänomene auf der Phrasenebene (Herr Meier und Frau Schmid > die beiden Lehrer/*Lehrerinnen). Soziopragmatische Aspekte werden anhand dialektaler Genus-Sexus-Diskordanzen vom Typ das Emma beleuchtet; auch syntaktische Serialisierungspräferenzen (Mann und Frau, Mama und Papa) variieren entlang sozio-pragmatischer Faktoren. Auf Diskurse rekurriert die Untersuchung zu typischen Argumentationsmustern gegen genderneutrale Sprache. Die Einstellungen nicht-binärer Personen zu gendersensiblem Sprachgebrauch spiegeln aktuelle sprachkritische Diskurse. Der Band ist eine hervorragende Informationsquelle für alle, die qualitativ hochwertige linguistische Argumentation und empirische Daten zu den komplexen Zusammenhängen zwischen Sprache und Geschlecht suchen.
Article
Full-text available
Two experiments tested whether the Dutch possessive pronoun zijn ‘his’ gives rise to a gender inference and thus causes a male bias when used generically in sentences such as Everyone was putting on his shoes . Experiment 1 ( N = 120, 48 male) was a conceptual replication of a previous eye-tracking study that had not found evidence of a male bias. The results of the current eye-tracking experiment showed the generically-intended masculine pronoun to trigger a gender inference and cause a male bias, but for male participants and in stereotypically neutral stereotype contexts only. No evidence for a male bias was thus found in stereotypically female and male context nor for female participants altogether. Experiment 2 ( N = 80, 40 male) used the same stimuli as Experiment 1, but employed the sentence evaluation paradigm. No evidence of a male bias was found in Experiment 2. Taken together, the results suggest that the generically-intended masculine pronoun zijn ‘his’ can cause a male bias for male participants even when the referents are previously introduced by inclusive and grammatically gender-unmarked iedereen ‘everyone’. This male bias surfaces with eye-tracking, which taps directly into early language processing, but not in offline sentence evaluations. Furthermore, the results suggest that the intended generic reading of the masculine possessive pronoun zijn ‘his’ is more readily available for women than for men.
Article
The aim of this study is to scrutinize Greenberg’s Universal 43, which predicts pronominal gender in the presence of nominal gender. On the basis of a sample of 500 gendered and ungendered languages, gender marking is examined in nouns, personal pronouns, possessors and possessums. Ungendered languages outnumber gendered languages. Eight out of 12 logically possible gender constellations are attested in the database. In keeping with Greenberg, languages with nominal gender show a strong bias towards gendered pronouns. There is a strong correlation between gendered personal pronouns and gendered possessors. Gendered possessums are cross-linguistically uncommon. The empirical patterns are brought about by a small set of theoretical principles. Gender is independently specified for each category. Gender marking is an effort. The strength of the correlation depends on the “distance” between two given gender sites. Coding gender twice in the same time frame creates a processing difficulty. Natural and grammatical gender conspire to generate the gender sensitivity of individual categories.
Article
Two experiments (N=48 each) were conducted to investigate gender-specific elements in the mental representation of short German texts. The texts contained a specific male or female designator (e.g., Mr. Smith, Mrs. Meyer) or a masculine generic phrase (CM) in either singular or plural usage (e.g., the student, the students) as text-subject. Two testphrases were constructed for each text, which did not appear in the text but reflected a masculine or feminine understanding of the text. Gender-specific associations were measured via the time that was required to reject the masculine and feminine test-distractors in a subsequent recognition task. Reading the texts with a specific male or female designator as text-subject increased the rejection time for the gender-congruent testphrases. For texts containing a GM as text-subject, the pattern of gender-specific associations was dependent on the grammatical numberof the GM-phrase. Reading a scenario containing a GM in the singular increased rejection times for the masculine test-phrases, while reading a scenario with a GM in the plural increased rejection times for the feminine test-phrases.
Article
The participants of this study, two hundred fifty French pupils aged fourteen and fifteen years, had to estimate their degree of self-efficacy toward various occupations. According to the experimental condition, occupations were presented only with the male grammatical gender [e.g., enseignant] or with the feminine grammatical gender [e.g., enseignant(e)]. Results obtained in this study indicate that, on average, pupils reported significantly more self-efficacy when occupations were presented with the feminine grammatical gender. Implications of this result are discussed with regard to the lack of the feminine grammatical gender in French for the most prestigious occupations.
Article
Current investigators of words, sentences, and other language materials almost never provide statistical evidence that their findings generalize beyond the specific sample of language materials they have chosen. Nevertheless, these same investigators do not hesitate to conclude that their findings are true for language in general. In so doing, it is argued, they are committing the language-as-fixed-effect fallacy, which can lead to serious error. The problem is illustrated for one well-known series of studies in semantic memory. With the appropriate statistics these studies are shown to provide no reliable evidence for most of the main conclusions drawn from them. A review of other experiments in semantic memory shows that many of them are likewise suspect. It is demonstrated how this fallacy can be avoided by doing the right statistics, selecting the appropriate design, and sampling by systematic procedures, or, alternatively, by proceeding according to the so-called method of single cases.
Article
Three experiments investigated how grammatical gender and gender stereotypicality influence the way person information is mentally represented. Participants read sentences about social groups denoted by nouns with different grammatical gender and stereotypicality. A following sentence contained a reference to the social group that qualified the group members as female, male, or neither one. Experiment 1 tested grammatically masculine nouns; Experiment 2 tested gender-balanced forms, composed of the masculine and the feminine or neither one; and Experiment 3 tested nouns without gender inflection. Stereotypicality varied within studies. Second sentence reading times differed depending on the fit between grammatical gender and stereotypicality of the first sentence’s subject and the subsequent information’s gender-relatedness. Both grammatical gender and stereotypicality contribute biological gender information to mentally represented person information. Strong grammatical input may override stereotypicality’s influence. The feminine’s influence seems to be weaker than the masculine’s. Results are discussed in the framework of the scenario mapping and focus approach.