Content uploaded by Alexandra Kate Frazer
Author content
All content in this area was uploaded by Alexandra Kate Frazer on Feb 25, 2016
Content may be subject to copyright.
Double Standards in Sentence Structure: 1
Running head: DOUBLE STANDARDS IN SENTENCE STRUCTURE
Double Standards in Sentence Structure: Passive Voice in Narratives Describing
Domestic Violence
Alexandra K. Frazer
Lehigh University
Michelle D. Miller
Northern Arizona University
Keywords: interpersonal violence; domestic violence; gender differences; verb voice;
language production; mass media
Double Standards in Sentence Structure: 2
Abstract
Previous research has shown that passive voice predominates in mass media reports
describing male violence against women. However, there has been little systematic study
of narratives describing female violence against men. We analyzed the impact of
perpetrator gender on verb voice, first in a content analysis of published news stories and
second in a new procedure for eliciting written narratives with male or female
perpetrators. Results revealed an increased frequency of passive voice when perpetrators
are male. These findings suggest that writers specifically prefer the passive voice to
describe male-on-female violence, rather than for violent or negative acts in general.
Double Standards in Sentence Structure: 3
When describing violent acts, writers and speakers make choices about sentence structure
that may reveal their underlying beliefs about these acts, such as whether the perpetrator
is solely responsible for the act, whether the victim is partly to blame, and how much
harm the victim suffered as a result. Particularly revealing is whether the writer chooses
an active or passive verb, as in the following example:
a. In the U.S., a man rapes a woman every 6 minutes.
b. In the U.S., a woman is raped by a man every 6 minutes.
(from Henley, Miller, & Beazley, 1995). Journalists tend to use the passive voice when
reporting male sexual violence against women, using passive voice by a ratio of more
than 2 to 1, according to a content analysis of news stories published in the Boston Globe
in 1981 and 1991 (Henley et al., 1995). This tendency is even stronger for other forms of
violent crime, with a passive-to-active ratio of over 3 to 1 for the verb murder (Henley et
al., 1995). By comparison, journalists tended to use the active voice for nonviolent
actions such as touched (passive-active ratio = .66) or thanked (passive-active ratio = .
37). Undergraduate student writers showed a similar pattern of increased passive voice
usage for sentences describing rape, a pattern that was particularly strong in writers with
greater acceptance of rape myths (Bohner, 2001).
Besides reflecting attitudes of writers, verb voice affects reader interpretations of
violent acts, particularly with respect to the victim’s role in bringing the crime on him or
herself. Psycholinguistic theories of the passive voice predict that readers will see the
object of the sentence – i.e., the crime victim – as more salient in passive voice sentences
(e.g., Clark, 1965; Clark & Begun, 1968; Johnson-Laird, 1968, 1977; Tannenbaum &
Double Standards in Sentence Structure: 4
Williams, 1968; Turner & Rommetveit, 1968). In theory, the passive voice would also
emphasize the causal role of the crime victim, relative to the perpetrator (Brown & Fish,
1983). Consistent with these predictions, both male and female readers show increased
acceptance of rape and battering of women after exposure to descriptions of sexual
assault written primarily in the passive voice, compared to readers exposed to active-
voice descriptions (Henley et al., 1995). Male readers attributed more victim
responsibility and less harm to victims after passive-voice descriptions, although female
readers did not show the same tendency (Henley et al., 1995; see also Bohner, 2001).
Similar effects of the passive voice also apply to non-sexual types of aggressive behavior
(Platow & Brodie, 1999). These effects of the passive voice on comprehension suggest
that writers’ use of verb voice influences how individual readers view violence against
women.
Why, then, do writers select active voice for some descriptions of violent acts and
passive voice for others? Gender of the people involved in the sentence is one important
factor, as men are generally seen as more active participants in the action being
described, while women are seen as passive recipients of the action (Deaux, 1976). This
may be more pronounced when the social interactants are a mixed gender pair (LaFrance,
Brownell, & Hahn, 1997). In assigning responsibility to the interactants of simple S-V-O
sentences, people tend to assign more responsibility to the subject of these sentences if
that person is male and the object of the action is female than vice-versa. Also, when a
woman is the recipient (object) of an action of a man (subject), the woman is more likely
to be seen as having triggered that action than in non-mixed gender pairs or if a woman
acts on a man (LaFrance et al., 1997). It is important to note that all of these were active
Double Standards in Sentence Structure: 5
voice sentences (S-V-O), so how people assign responsibility to passive voice sentence
interactions is more unclear, especially because it has been hypothesized that with regard
to action verbs, the person who is interpreted as causal is the sentence subject (Brown &
Fish, 1983).
However, this tendency to portray men as active agents may be offset by a
predisposition to play down men’s responsibility for violence against women (Penelope,
1990). Further complicating the picture is the fact that in mass media reports of violence,
men are overwhelmingly the perpetrators, making it impossible to know whether writers
choose verbs that play down men’s causal role as perpetrators of violence, or that play
down the causal role of perpetrators in general.
Bohner (2001) favors a somewhat different explanation, suggesting that writers
use the passive voice to psychologically distance themselves from disturbing, violent
actions such as sexual assault. In Bohner’s procedure, participants who were more
accepting of rape myths used a higher proportion of passive voice verbs to describe a
rape shown in a film excerpt, compared to participants who were less accepting of rape
myths. These results do help establish a connection between writers’ beliefs about
violence and their propensity to use the passive voice. However, given that all the
perpetrators in these film clips were male, the question remains whether writers use the
passive voice to de-emphasize the role of men as perpetrators of violence, or to de-
emphasize perpetrators of violence in general.
One way to address this question is to compare descriptions of violent acts in
which women are the perpetrators with those in which men are the perpetrators. If writers
are generally averse to using the active voice to describe violent acts, there will be similar
Double Standards in Sentence Structure: 6
proportions of passive voice usage for woman-on-man and man-on-woman violence. If,
on the other hand, writers use passive voice specifically to play down men’s causal role
in committing violence against women, there will be a decreased proportion of passive
voice in descriptions of woman-on-man violence. We focused on domestic violence –
battering one’s spouse or romantic partner, or other physical assault up to and including
murder – because it is committed with some frequency by women against men (although
the reverse is much more common as well as typically more severe; Tjaden & Thoenes,
2000).
Study 1
Using a procedure modeled after Henley et al. (1995), we searched for newspaper
articles describing domestic violence in heterosexual couples. We then analyzed the
prevalence of passive voice in sentences specifically referring to the violent acts that
were the subject of these articles. Our hypothesis was that passive voice would be more
prevalent in sentences describing male-on-female violence.
Method
Materials. We searched the Boston Globe for articles published in 2003 and 2004
about severe domestic violence. We chose the Boston Globe because it was the
newspaper used in the Henley et al. (1995) content analysis and has a searchable database
spanning a number of years. To identify articles which could potentially be about
domestic violence and could have either gender as the perpetrator, we selected any whose
lead paragraph contained the words “death”, “murder”, “kill”, “husband”, “wife”, “man”,
and “woman.” This search yielded 200 lead paragraphs for subsequent analysis.
Coding. Two coders (S.K. and J.H) completed a training set of materials with an
Double Standards in Sentence Structure: 7
interrater reliability of 100% before coding the study materials. Coding consisted of
identifying each verb that referred directly to the violent act that was the subject of the
story, and then coding that verb for voice (active vs. passive), gender of the perpetrator,
and gender of the victim.
Results
Not all of the 200 lead paragraphs contained sentences that were appropriate for
the coding procedure; some did not address the actual violence in the lead paragraph,
others were not about specific cases of domestic violence. One verb phrase was excluded
from analysis because the victims it described were a mixed-gender group and one was
excluded because the gender of the victims was not clear from the lead paragraph. Of the
remaining 53 verbs coded, 48 described male-on-female violence and 5 described female-
on-male violence. There were 13 passive and 35 active verbs in the male-on-female
condition, compared to 1 passive and 4 active in the female-on-male condition; i.e., verbs
were passive 27.1% of the time when men were the perpetrators, but only 20% of the
time when women were the perpetrators. The lack of female-on-male verbs ruled out
further statistical analysis of this trend.
Discussion
There was a trend toward increased use of the passive voice when men were
perpetrators and women victims, but this finding must be interpreted with caution
because of the small number of stories describing cases where the woman is the
perpetrator. Our search turned up only 5 verb phrases describing this type of violence in
lead paragraphs covering 2 years of reporting in a metropolitan newspaper, as compared
to the 48 we found for violence committed by men against their spouses. This disparity
Double Standards in Sentence Structure: 8
may be an inherent limitation of mass media reports as the subject of content analysis
about inter-gender violence.
Mass media reports are also impractical for examining the relationship between
gender of the writer and description of violent acts, given that it is difficult to establish
the gender of a story’s primary writer as well as the gender of any editors or coauthors.
This relationship is important for shedding light on the underlying mechanisms for use of
the passive voice to describe inter-gender aggression, as men and women presumably
have different views of important factors such as their identification with the victim
versus perpetrator, perceived likelihood of being the victim or perpetrator of aggression,
and motivation to psychologically distance oneself from the aggressive act.
We designed Study 2 to address the inherent limitations of using published mass
media reports to study the relationship between gender, verb voice and violence. Our goal
was to generate equal numbers of female-on-male as male-on-female stories, as well as to
track the influence of writer gender on the syntactic structure of violence descriptions.
Shifting our approach from analysis of existing stories to eliciting new ones introduces a
new set of challenges, as production is notoriously difficult to study using controlled
experimental techniques (e.g., Bock & Griffin, 2000).
To deal with these challenges, we presented participants with highly simplified
descriptions of domestic violence scenarios and asked them to incorporate these in
written narratives. We then analyzed these elicited narratives for the relative frequency
of active and passive voice verbs. The scenarios were systematically varied so that there
were equal numbers of male and female perpetrators, allowing us to contrast the
proportion of passive voice usage for male versus female perpetrators, while holding
Double Standards in Sentence Structure: 9
constant the nature of the violent act as well as other details of the scenario. The
scenarios were not described in narrative format, as the verb voice used in text
descriptions could affect the verb voice used in participants’ subsequent elaborations
through syntactic priming (Bock, 1986a). Instead, we presented the parameters of each
scenario in a matrix that listed information the type of event, victim (name and age),
perpetrator (name and age), weapon, date, and place.
Study 2
Method
Participants. Participants were 86 Northern Arizona University undergraduate
students enrolled in an Introduction to Psychology course. Participants were 18 to 25
years old (mean age =18.52 years, SD=1.03). Participants received extra credit in their
Introduction to Psychology course for participating in this study.
Materials. We presented participants with an information grid that provided
information regarding the event, victim (name, gender and age), perpetrator (name,
gender and age), weapon, date, and place. Figure 1 shows an example grid. Names of
victim and perpetrator were gender-specific. Each participant viewed two grids, one for
the target event, and another filler crime event. Materials were counterbalanced so that
for the target event, half of the grids specified a male victim and female perpetrator, and
half of the grids specified a female victim and male perpetrator. Participants also filled
out a brief demographic questionnaire.
Procedure. Written instructions stated that the study topic was how students
write about crime, and asked participants to write two 50- to 100-word stories that
incorporated the information shown on the grids.
Double Standards in Sentence Structure: 10
Results and Discussion
Four participants were excluded for deviating from the instructions: One wrote an
opinion about the event, and three wrote stories in which the weapon was a gun rather
than the knife specified in the grid. One experimenter (A.F.) read each critical story and
recorded each instance of a “target sentence,” defined as any sentence containing a verb
which directly described the domestic violence. Some participants generated more than
one target sentence, for a total N of 104 target verbs. Each sentence was then coded as
active or passive, with passive defined as any form of the verb “to be” + the past
participle (e.g., was beaten, had been killed).
Table 1 shows all target sentences broken down by participant gender, verb voice
(active vs. passive) and condition (male-on-female violence vs. female-on-male
violence). Participants produced a significantly higher proportion of passive voice
sentences for the male-on-female violence condition, x2=3.917, p=.048. As Table 1
shows, the trend toward increased passive voice usage in the male-on-female condition
was present for both female and male participants.
Examination of the elicited stories revealed an unanticipated subsidiary finding
that participants often included a non-elicited justification for the violence in their stories.
After coding stories for the presence or absence of a justification, we found that there was
a significant relationship between condition and presence of a justification , x2=14.595,
p<.001. Specifically, participants were more likely to include a justification when
describing female-on-male spousal violence than when describing male-on-female
spousal violence. For the male-on-female condition, 6 stories included a justification and
44 did not. For the female-on-male condition, 29 stories included a justification and 25
Double Standards in Sentence Structure: 11
did not. These justifications generally manifested themselves in later sentences in the
narrative rather than being part of the sentences analyzed for verb voice. Clearly,
participants did see the abuse differently when it was a woman perpetrator, as evidenced
by more justifications, but the stories still described violence. 1
General Discussion
Our results do not support the idea that writers use passive voice to describe
violence against women simply because of a general aversion to using active voice to
describe violent acts. Rather, there appears to be a more complex relationship involving
gender of the people involved in interpersonal violence, such that when women are the
perpetrators and men the victims, active voice becomes more common. This trend,
though difficult to observe using content analysis of actual published news reports,
emerged in our elicited narrative paradigm. Differences in how writers view female-on-
male versus male-on-female violence are evident in the (probably) unconscious process
of structuring sentences. Specifically, female attackers’ causal role may be more salient
than that of male attackers, resulting in descriptions that highlight women’s responsibility
and de-emphasize men’s responsibility for acts of interpersonal violence. Furthermore,
participants may form more elaborate explanations about female-on-male violence, as
evidenced by the increased number of spontaneously generated justifications in the
female-on-male condition. This finding – that third-party writers made excuses and
justifications for female perpetrators – is an interesting complement to the finding that
when male perpetrators of domestic violence talk about their actions, they tend to use
language that diminishes their responsibility through making excuses and justifications
(O’Neill & Morgan, 2001).
Double Standards in Sentence Structure: 12
Also, Bock (1986b) has demonstrated that normal sentence production utilizes an
abstract representation that includes the semantic information about the information
contained in the sentence. This abstract representation is influenced by the semantic
information available; therefore, semantic characteristics can directly influence the syntax
of the sentence. In the current study, this means that semantic characteristics related to
“male” and “female” may have influenced the selection of active and passive voice. Bock
also made note of the fact that the passive voice is used relatively little by English
speakers, especially when the sentence contains a human agent and a human patient.
Taken together, this suggests that the semantic characteristics of men and women may
differ with regards to violent behavior, resulting in increased passive voice usage for
female-on-male violence.
It is also important to consider how participants represent victims and
perpetrators, not just men and women. Linguistic devices or operations, such as the
passive voice, may be used in a variety of ways to manipulate the representations of the
violent events, and as such, are able to be used to conceal violence, reduce the
responsibility of perpetrators, and even blame the victim for the violence (Coates &
Wade, 2007). In doing this, third-parties are constructing misleading accounts of
violence. As this relates to the passive voice, it seems that it is one way of both reducing
the responsibility of the perpetrator, and blaming the victim for the violence, by placing
the victim in what is generally the causal role of the sentence, the subject (Brown & Fish,
1983).
Previous studies have found asymmetries in passive voice usage for the domain of
sexual assault by men against women; our findings extend to a new domain, domestic
Double Standards in Sentence Structure: 13
violence, in which gender roles can be more easily reversed and which is not specifically
sexual in nature. This asymmetry may extend to a number of other types of acts for
which writers wish to de-emphasize the responsibility of the agent, for example, de-
emphasizing their own role in making a mistake (Penelope, 1990). It should be noted,
however, that Henley, Miller, Beazley, Nguyen, et al. (2002) found no increase in passive
voice in news reports of violence by heterosexual individuals against gay men and
lesbians, another arena in which writers may be motivated to emphasize the role of the
victim over that of the perpetrator. While the present results suggest that the findings
from the domain of sexual assault do generalize to other actions, there may be subtle
differences in attitudes that affect passive voice usage across these other domains. These
differences could be explored in future research.
Future research can also examine the pattern of passive voice usage when the
nature of the interpersonal act changes altogether, such as for positive acts. For example,
in the case of the verb save, is the passive voice more common when a woman saves a
man, versus when a man saves a woman? Such an investigation could shed light on
whether writers’ tendency to de-emphasize male agency disappears when the action is
positive, and whether it reverses when a woman is the subject of the sentence. Henley et
al. (1995) found that the passive voice was relatively rare for verb thanked, suggesting
that readers emphasize the subject in sentences describing positive behavior. However,
these results were not broken down by the gender of the people involved in the action,
leaving open the question of whether gender-related attitudes affect descriptions of
positive actions as well as the negative ones examined here and previously. Preliminary
follow-up research by the authors has indicated that regardless of whether positive
Double Standards in Sentence Structure: 14
actions that are typically gender specific (e.g. a marriage proposal) are male-on-female,
or female-on-male, the active voice was used almost exclusively (Frazer, 2007). This
lends credence to our supposition that the passive voice usage for violence against
women is not simply a result of the interaction being male-on-female.
Turning now to questions of gender differences among writers, our results
replicate and extend Bohner’s (2001) finding that men and women are similar in how
they use passive voice to describe interpersonal violence. These findings are somewhat
surprising in that men and women presumably have different views about interpersonal
violence, given that women are far less likely to perpetrate this type of act, are less
accepting of rape myths (Muir, Lonsway, & Payne, 1996) and are less likely to blame
victims of domestic violence (Bryant & Spencer, 2003). The findings also suggest an
asymmetry in production versus comprehension of passive voice, as men were more
affected by reading passive voice, compared to women (Henley et al., 1995).
The effect of the passive voice on comprehension bring us back to the wider
social significance of patterns of verb voice in descriptions of domestic violence.
Structuring sentences in a way that emphasizes women’s causal role in such violence,
while de-emphasizing men’s role, may be one of several mechanisms by which writers
and speakers express their attitudes about gender, sex, and power. These mechanisms
include the use of erotic rather than violent terminology in rape trials (Bavelas & Coates,
2001), passive voice use in rape descriptions (Bohner, 2001; Henley et al., 1995) and
victim-blaming language in descriptions of rape (Kanekar, Kolsawalla, & D’Souza,
1981). Such expressions may in turn shape whether readers and listeners interpret these
acts as voluntary acts of violence against an undeserving victim, or as unfortunate
Double Standards in Sentence Structure: 15
experiences that women – at least in part – bring on themselves.
Double Standards in Sentence Structure: 16
Authors’ Note
Portions of this research were submitted to satisfy the master’s thesis requirement for
Alexandra Frazer. We are grateful to Robert Till and Derrick Wirtz for helpful
comments about the design and results, and our two research assistants, Stefanie Burnett
Keefe and Jerome Hamman, for coding the Study 1 materials. We also thank Howard
Giles and two anonymous referees for their helpful comments on an earlier draft of this
article. Portions of this research were presented at the April 2005 meeting of the Rocky
Mountain Psychological Association in Phoenix, Arizona. Correspondence concerning
this article may be directed to Michelle D. Miller, Department of Psychology, Northern
Arizona University, Flagstaff, AZ 86011-5106
Double Standards in Sentence Structure: 17
Note
Example female-on-male with justification narrative: “There was a brutal murder
last night at the Smith residence. Joe Smith’s, husband of Amanda Smith, life was cut
short. The unbelievable happened. His wife of 8 years stabbed him to death in their living
room with a kitchen knife while he was taking a nap. She claims that he was cheating,
and when confronted, he lied. She wasn’t going to harm him, she just wanted to scare
him, but he attacked first so she had no choice. But the repetitive stab wounds tell a
different story.”
Double Standards in Sentence Structure: 18
References
Bavelas, J., & Coates, L. (2001). Is it sex or assault? Erotic versus violent language in
sexual assault trial judgments. Journal of Distress and the Homeless, 10, 29-40.
Bock, J. K. (1986a). Syntactic persistence in language production. Cognitive
Psychology, 18, 355-387.
Bock, J. K. (1986b). Meaning, sound, and syntax: Lexical priming in sentence
production. Journal of Experimental Psychology: Learning, Memory, and
Cognition, 12, 575-586.
Bock, J. K., & Griffin, Z. (2000). Producing words: How mind meets mouth. In L.
Wheeldon (Ed. ), Aspects of language production (pp. 7-47). Philadelphia:
Psychology Press.
Bohner, G. (2001). Writing about rape: Use of the passive voice and other distancing text
features as an expression of perceived responsibility of the victim. British Journal
of Social Psychology, 40, 515-529.
Brown, R., & Fish, D. (1983). The psychological causality implicit in language.
Cognition, 14, 237-274.
Bryant, S. A., & Spencer, G.A. (2003). University students’ attitudes about attributing
blame in domestic violence. Journal of Family Violence, 18, 369-376.
Clark, H. H. (1965). Some structural properties of simple active and passive sentences.
Journal of Verbal Learning and Verbal Behavior, 4, 365-370.
Clark, H. H., & Begun, J. S. (1968). The use of syntax in understanding sentences.
British Journal of Psychology, 59, 219-229.
Coates, L., & Wade, A. (2007). Language and violence: Analysis of four discursive
Double Standards in Sentence Structure: 19
operations. Journal of Family Violence, 22, 511-522.
Deaux, K. (1976). Sex: A perspective on the attribution process. In J. Harvey, W. Ickes,
and R. Kidd (Eds.), New Directions in Attribution Research (pp. 335-352).
Hillsdale, NJ: Lawrence Erlbaum Associates.
Frazer, A. K. (2007). Positive and negative events, gender, and the use of the active and
passive voice. Unpublished master’s thesis, Northern Arizona University,
Flagstaff, Arizona, USA.
Henley, N. M., Miller, M. D., & Beazley, J. A. (1995). Syntax, semantics, and sexual
violence: Agency and the passive voice. Journal of Language and Social
Psychology, 14, 60-84.
Henley, N. M., Miller, M. D., Beazley, J. A., Nguyen, D. N., Kaminsky, D., & Sanders,
R. (2002). Frequency and specificity of referents to violence in news reports of
anti-gay attacks. Discourse and Society, 13, 75-104.
Johnson-Laird, P. N. (1968). The interpretation of the passive voice. Quarterly Journal
of Experimental Psychology, 20, 69-73.
Johnson-Laird, P. N. (1977). The passive paradox: A reply to Costermas and Hupet.
British Journal of Psychology, 68, 113-116.
Kanekar, S., Kolsawalla, M. B., D’Souza, A. (1981). Attribution of responsibility to a
victim of rape. British Journal of Social Psychology, 20, 165-170.
LaFrance, M., Brownell, H., & Hahn, E. (1997). Interpersonal verbs, gender, and implicit
causality. Social Psychology Quarterly, 60, 138-152.
Muir, G., Lonsway, K.A., & Payne, D. L. (1996). Rape myth acceptance among Scottish
and American students. Journal of Social Psychology, 136, 261-262.
Double Standards in Sentence Structure: 20
O’Neill, D., & Morgan, M. (2001). Pragmatic post-structuralism (I): Participant
observation and discourse in evaluating violence intervention. Journal of
Community and Applied Social Psychology, 11, 263-275.
Penelope, J. (1990). Speaking freely. New York: Pergamon Press.
Platow, M. J., & Brodie, M. (1999). The effects of linguistic voice on evaluations and
attributions of ingroup and outgroup members. Asian Journal of Social
Psychology, 187-200.
Tannenbaum, P. H., & Williams, F. (1968). Generation of active and passive sentences
as a function of subject or object focus. Journal of Verbal Learning and Verbal
Behavior, 7, 543-548.
Tjaden, P., & Thoenes, N. (2000). Prevalence and consequences of male-to-female
and female-to-male partner violence as measured by the national violence against
women survey. Violence Against Women, 6, 142-161.
Turner, E. A., & Rommetveit, R. (1968). Focus of attention in recall of active and
passive sentences. Journal of Verbal Learning and Verbal Behavior, 7, 543-548.
Double Standards in Sentence Structure: 21
Table 1
Target Sentences Broken Down by Participant Gender, Verb Voice and Gender of
Perpetrator/Victim
Gender of perpetrator/victim Participant Gender Active Passive
N N
Male perpetrator/female victim Male 4 11
Female 15 20
Total 19 31
Female perpetrator/male victim Male 8 7
Female 23 16
Total 31 23
Double Standards in Sentence Structure: 22
Example Information Grid Presented To Elicit Narrative Descriptions Of Domestic
Violence Scenarios
Event Death
Victim Joe Smith (male, 31 years old, husband)
Perpetrator Amanda Smith (female, 33 years old, wife)
Weapon Knife
Date June 9th, 1996 at 7:18 pm
Place The Smiths’ residence; Anytown CO
