Gender and Language Processing
December 26, 2019
So let’s talk about gender. It’s a construct
that’s a large part of our lives, and we interact with it in a whole bunch of different ways.
But even for those of us who think that we’ve gotten beyond gender considerations, there’s
still a lot of unevenness in how our brains handle the language side of things.
Down to the deepest parts of our minds, we’re still fighting our preconceptions. I’m Moti
Lieberman, and this is the Ling Space. So there are two different concepts that linguists
talk about when it comes to gender. The first of these is called grammatical gender. This
is an abstract concept that gets applied to nouns, and then matched on the adjectives
and verbs around them. A whole slew of languages around the world
use gender marking, from French to Polish to Hebrew to Tamil. This sort of gender isn’t just assigned
to people, either. Every noun in your language will get it, even those for ideas and objects.
Like, in French, you get a masculine ‘un miroir’ for a mirror, but a feminine ‘une épée’ for
a sword. This sort of gender doesn’t always match
across languages, either. So, in German, a sword isn’t feminine, but it’s neuter
instead, like ‘das Schwert’. For people, though, the terms and genders
usually line up the way we expect. Let’s look at Irish: the word for boy, buachaill,
is masculine, and the word for woman, bean, is feminine.
But even that’s not always true: the word for girl, cailín, turns out to be masculine,
as well. Grammatical gender is an important concept,
and there’s been a lot of research done on it:
the variety of gender systems in languages, and the acquisition of gender, and even how
grammatical gender can actually guide the interpretation of sentences. And I promise,
we’ll come back to talk about it more in the future, but for today, we’re going to
focus on the regular, more commonplace meaning of gender. So, like it or not, pretty much every society
has culturally conditioned ideas of what qualifies as masculine or feminine.
Depending on where you’re from, you might have different expectations about appropriate
behavior for men or women, but you’ll still have something.
We use those beliefs for a lot of things, and one of them is to help us quickly build
meanings as words are hurled our way like a series of rapid fireballs. Our thoughts about gender influence the way
that we search for our words and build our sentences. And when you look at how people’s
brains are doing this, it turns out that the thoughts that we have deep down aren’t always
the most progressive ones. Take, for example, a 2009 neurolinguistic study that wanted to
see whether stereotypes about gender reached all the way down into word associations. The researchers put together a study using
event-related brain potentials, or ERPs, that we talked about back in Topic 18.
In particular, they wanted to look at whether associating men or women with words that match
or don’t match stereotypical gender views caused an N400, the brain’s way of saying,
“Wait, what was that?” When we talked about the N400 before, we discussed
how it showed the cost of bringing new, unexpected words into a sentence, even if they were grammatically
correct. So the last word of a sentence like “Terry kicked the enemy in the chest”
would be easier to process than “Terry kicked the enemy in the goose.” Goose might be grammatically correct there,
like, it’s still a noun, but it’s not an appropriate noun to use. So our poor neurons
have to work extra hard to process the sentence, and that shows up as an N400.
But it’s not just sentences where we can find evidence of our brains chugging away.
We can also find similar N400 effects for word pairs. Our brains have an easier time matching up
things that are related to each other, like “tournament” and “champion,” than
things that don’t have any apparent relationship at all, like “ash” and “modest.” But
a pair like tournament and champion have an unbreakable tie: if you hold a tournament,
you’ll have a champion. When it comes to gender, though, there’s
no logical tie that just can’t be broken. The researchers in the study wanted to see
how their participants’ brains would deal with processing words that are more associated
with men or women. To do this, for each item in the study, they’d
first show the word “man” or “woman,” and then present a second word that’s stereotypically
linked with one gender or the other. Some of these were traits, like ‘nurturing’
or ‘aggressive’, and some were nouns, like ‘makeup’ or ‘mechanic’. And then the participants
were asked to indicate whether the second word they heard matched the gender category
of the word before it. So a round would work like this: the participant,
with the brain-measuring electrode cap on their head, would first see “man,” and
then “secretary”, and then they would push a button to show whether they thought
this was an appropriate gender match. Now, again, it’s important to note, there’s
nothing concrete tying men and women to any of these terms. Men can wear fancy make-up
and be caring, and women can be aggressive, tough mechanics.
But we have our socially gendered perception about who these terms should go with, and
it’s pretty strong. The researchers in the study found that when
the words matched gender stereotypes, like ‘woman’ and ‘doll’, participants answered
quickly and matched them easily. When the words didn’t match a stereotype,
like ‘woman’ and ‘cigars’, they were slower and more unsure about matching the gender
expectations. And this carried over to their brains: the N400 effect was much stronger
for the non-matched pairs. Participants really had a harder time linking
up ‘woman’ and ‘powerful’ than ‘woman’ and ‘gossipy’. So we can find the influence of gender on
how we process words, and even in how our brains react. But we just don’t really normally
encounter words in isolation. And people in that study were really primed
to think about gender, I mean, they were explicitly asked to do just that.
What happens when people are dealing with full sentences? How does gender influence
you then? To answer that question, let’s take a look
at a 1997 neurolinguistic study. This time, the researchers used the power
of reflexive pronouns to look at how gender influenced interpretation.
Reflexives are items like himself and herself, and they’re really demanding little terms.
In order to use a reflexive grammatically, it has to be associated with something in
the sentence. You can’t just have a sentence like “Himself
fought lots of his clones.” Himself needs to be somebody, or no fighting can happen.
But more importantly for our purposes, it’s not enough that there be some noun just be
there to go with himself or herself. It has to be the right gender, too. If we have a
sentence like “Billy hit himself with the cane,” that’ll only be good if Billy is
male. So how can we use these facts about pronouns
to look at gender bias? Well, here’s the really clever bit: some nouns, by definition,
have to go with one gender or the other. Nouns like “bachelor” and “king” must
be male, and things like “niece” and “nun” must be female. That’s just part of their meanings.
But other nouns are only stereotypically connected to one gender or the other, like we saw before. It’s not part of the definition of “police
officer” or “pilot” that they be male, or for “cheerleader” or “model” to
be female. We just have gender biases one way or the other.
So now, let’s bring the reflexives back into the mix.
If you have something that, by definition, has to be male or female, then putting in
the wrong gender reflexive just makes the whole sentence bad. You can’t say “The
famous ballerina prepared himself for the performance.” Ballerina and himself don’t match, and no
matter what kind of stereotypes you have, there’s nothing you can do to rescue this
sentence. If it’s not ‘herself’, it’s a mistake. But if you have something that just
doesn’t match the stereotype, what then? Take a sentence like “The bouncer taught
herself to identify suspicious behavior.” Now, that’s a totally fine, grammatical
sentence of English. The only thing that it requires of you is
that you overcome your gender bias so you can associate ‘bouncer’ with ‘herself’, and you’re
set. But do people do that? To figure this out, participants were hooked up to the ERP
machine to see what was going on with their brains.
Then they read sentences where the reflexives either matched or didn’t match for both
the definitional and stereotypical conditions. At the end of each sentence, they were asked
to judge whether what they’d just seen was an acceptable sentence of English. So what did
they find? Well, just looking at the participants’ judgments, the researchers found significant
differences for both cases. So whether they were driven by grammatical
need or gender stereotype, they still found the matches significantly better than the
mismatches. So “The bouncer taught herself” is always
more unacceptable than “the bouncer taught himself.”
The size of the effect was stronger for the ones where the noun really needed it by definition,
but both trends were pretty solid. But let’s step back for a second. “Unacceptable”
can mean a lot of things. It could mean grammatically bad, like attaching ballerina to himself.
Or it could mean kinda weird from the stereotypical point of view, like putting beautician with
himself. Maybe people just had a harder time associating a reflexive with a noun that it
doesn’t usually go with, and that’s why they called it unacceptable. They did still find 77% of the stereotypical
mismatch cases okay in the end, too. So it’s not like they were rejecting female
bouncers all the time, it’s just that they liked them reliably less than the male ones.
If that’s what’s going on in here, it should show up in the ERP data as a semantic
mismatch, so the N400. That’s what the N400 does – it shows you
when you have a hard time fitting a word that’s grammatically okay into the sentence. Except… this time, that’s not what the
researchers found. Instead, they found a different ERP effect, a P600, in both mismatch cases.
But the P600 is associated with syntactic problems, not semantic ones. It’s what you
get in a sentence like “Shen eat all the crabs”.
That’s understandable for the cases where by definition, the words just can’t work
together, like ‘niece’ and ‘himself’. But it means that rather than seeing ‘soldier’
plus ‘herself’ or ‘gymnast’ plus ‘himself’ as okay, but weird, the initial brain reaction was
just “This is bad. This sentence is just ungrammatical”. Now, we can get past these readings. Even
with the brain saying “whoa, this is wrong” at first, they still eventually accepted the
stereotypical mismatch over three-quarters of the time.
And no one is saying that women can’t be powerful, and men can’t be nurturing. But
the findings from neurolinguistics research so far pretty strongly tell us something that
we may not want to hear. When it comes to gender, our first interpretations are unfortunately
predictable. So we’ve reached the end of the Ling Space
for this week. If you overcame your initial preconceptions, you learned that there are
two kinds of gender in language, grammatical and sociological; that gender stereotypes
underlie our word associations, whether we want them to or not; and that in processing
sentences, gender biases temporarily keep us from considering some legitimate interpretations. The Ling Space is produced by me, Moti Lieberman.
It’s directed by Adèle-Elise Prévost, and it’s written by both of us.
Our editor is Georges Coulombe, our production assistant is Stephan Hurtubise, our music
and sound design is by Shane Turner, and our graphics team is atelierMUSE. We’re down
in the comments below, or you can bring the discussion back over to our website, where
we have some extra material on this topic. Check us out on Tumblr, Twitter and Facebook,
and if you want to keep expanding your own personal Ling Space, please subscribe. And
we’ll see you next Wednesday. Ne sta eva hose voomatse!