Laudre (laudre) wrote in linguaphiles,
linguaphiles 5 Surprising Ways Your Language Affects How You Think

A listicle, as described in the subject line.

The article implies or outright states a Sapir-Whorf perspective on interpreting the studies, even some that manifestly aren't (e.g. prosody), but behind that is some interesting tidbits on various research findings over the years.

  • Post a new comment


    Anonymous comments are disabled in this journal

    default userpic

    Your reply will be screened

    Your IP address will be recorded 

I can't talk about the other things, but I don't agree with his opinions about gendered languages. I'm Spanish and the fact that we say, i.e., "la mesa" and "el puente" doesn't mean we think of the table as female or the bridge as male.
I'm struck by the fact that some of the studies have references to specific experiments, while others link only to correlational relationships --- including gender*, and also whether a language has a future tense. The experiments seem legit to me, but the correlational studies are problematic. There's so much data about linguistic features and cultural properties that it's really easy to find absolutely spurious associations that are simply statistical accidents. Here's a paper about that:

Roberts, Seán and James Winters. 2013. Linguistic Diversity and Traffic Accidents: Lessons from Statistical Studies of Cultural Traits. PLOS One DOI: 10.1371/journal.pone.0070902.

Linke here:

*It's really hard to sort out what is made-up garbage by the blogger, and what has actually been proposed. For example, bridges are feminine in German because you walk over them? I suspect that German gender came into its modern form well before that particular metaphor became extant.


January 13 2014, 19:32:50 UTC 4 years ago Edited:  January 13 2014, 19:41:22 UTC

The first part, about grammatical gender, was almost enough to discourage me from reading the rest.
The rest appears to be less unrealistic. I've read some articles on how children learn intonations before learing actual words; seems hardly surprising.
I've also read some research on how thinking in foreign language makes you more logical, but the results were inconclusive, iirc. And it's not about languages per se, so much as about our minds using a non-native operating system. Same with switching opinions when switching to a different language, probably.
Yeah, the gender bit at the beginning had me rolling my eyes.

The other things, I can see some sort of reasoning, but nothing that requires a Whorfian interpretation. I do know that, for example, whenever I talk about certain particularly redneck topics (e.g. guns, big old cars with big engines), I'm prone to shift, involuntarily, to what, in NC, is called a "country" accent (i.e. a Southern drawl). I think that has a lot more to do with context, though -- that is, the people I'm most likely to discuss these topics with are most likely also going to speak in such a dialect. That is, I'm pretty sure what we're seeing in that case are the results of social cues that are linked to a particular language, dialect, and/or mode of speech.

Similarly, I suspect that the "more logical" thing isn't because the language itself forces you to be more logical, but because when you're responding to a question in a non-native language, you're engaging more active analysis in your cognition just to interpret the question, which affects how you analyze the question itself -- i.e., more logically and analytically, less intuitively. Also, I wonder about the sequencing, and about participation -- are you asking the same question of the same people at different times, in different languages? If so, are you giving the question in the same language each time (i.e. native language first, L2 second)? That seems like something you'd want to control for; i.e., have some people participate in only one language and some in both languages, and have the order randomly chosen, and with different gaps, and so on.
Exactly. Such experiments are tricky, you'd need to take into account a lot of unrelated matters to make sure your results are valid. It's an interesting topic, and I actually like the Whorfian theory, if not taken to extremes. But I haven't seen any convincing results yet.
I've just clicked through to read the study in question, and it seems well controlled. Four different languages were studied: English speakers in Japanese, English speakers in French, and Korean speakers in English. There were over a hundred people in each group, and they were randomly assigned to perform the task either in their native language or their non-native language, and they were randomly assigned to either the "loss-frame" or the "gain-frame"---so there were no ordering issues to control. Each person only saw one version, in one language.


January 13 2014, 19:48:27 UTC 4 years ago Edited:  January 13 2014, 19:56:16 UTC

Just to clarify, on the first matter.

1. Grammatical gender has nothing to do with meaning, at least not in German. Or Russian. Or Latin. Or any other Indo-European language I know. It's a simple matter of what the given word ends with. If it ends with -x, it follows this inflection paradigm. If it ends with -y, it follows that one. These categories are traditionally called "genders" in Indo-European linguistics, although they might as well have been called No.1, No.2, etc.

2. I'm no expert on statistics, but according to the article's logic, ancient China, for example, should have been the least gender discriminating country of all times. Neither Chinese or, say, Japanese has any gender markers, nor had them - ever. Unless we count descriptive constructions like 女の子 = female+Gen+child = girl. Never prevented medieval gender discrimination in the region.
(I think I'm just being a devil's advocate here; feel free to ignore my quibbles, but it's such a fun argument!)

1. Oddly enough, Lera Boroditsky has done some really interesting work on German and Spanish, showing that there do seem to be some male/female associations with grammatical gender. She found that grammatically feminine items were more "female," and controlled for inherent properties by taking pairs of items that were masculine in German and feminine in Spanish (and vice versa), like "bridge" and "key." She also replicated these effects in an artificial-language learning experiment, in which she showed that it is probably due to the association between grammatical gender and real-world gender. Here's a link to some of her work:

2. (I use statistics in all of my research.) A correlation does not have to be perfect. Just because a trend exists doesn't mean that every data point corresponds to it. It is possible that China and Japan are simply outliers, or perhaps a combination of other factors might overwhelm the effect of non-gendered nouns in Chinese and Japanese.

You do, however, make an excellent point about the importance of time period. I think almost every society that has a written history would show enormous gender discrimination, regardless of the grammatical system that the language employs. Since gender systems remained (more) constant that just about every other aspect of their culture over time, how likely is it that the language's gender system is responsible for modern differences in gender equality, rather than anything else about culture that has changed over the years? I have by far the most difficulty believing this claim over all the others (except perhaps the association between future tense and whatever else it was.)
1. Thanx for the link, I'll look into it.
I understand that people may associate grammatical gender with actual gender. But not to the extent described in the article referenced by the OP. I mean, okay, if it's a bridge as a cartoon character, it'd be male in a Russian cartoon and female in a German one. But other than that...
Grammatical gender for "human/person" is masculine in Russian and feminine in Ukranian. Should we take it as a sign of female and male discrimination in Russia and Ukraine, respectively?
Also, in Slavic languages we have the so-called "general gender" nouns that may apply to any gender while being grammatically "masculine" or "feminine". Most of them give positive or negative description, like молодец (m) = "well done!", "good girl/boy!"; or сволочь (f) = bastard/jerk. These words (and their attributes in a sentence) usually have clear gender markers with no regard to the actual gender of the person described. Does it prove that being a smart cookie (умница) is intrinsically feminine in Russian? Or that a girl calling her boyfriend "sunshine" (солнце, neutral) doesn't see him as a male and a partner? I somehow doubt it :>

2. That's exactly my point.
But not to the extent described in the article referenced by the OP

Well, yes. That's generally what happens when popular press gets hold of scientific findings. I remember reading a newspaper article that touted a discovery of the cure for cancer in the headline, and then in the last column they have a quote from the researcher, saying, something like, "yeah, this is pretty cool! We think in five or ten years we might have a prototype to test on mice!"
I think Keith Chen's claims have been debunked to my satisfaction by the Language Log team. As one of the authors of the EUROTYP survey (Östen Dahl) points out in the comments here, Chen imposed a dubious binary classification on the data which wasn't in the source material and then used that as the basis for his hypothesis. That seems typical to me of what happens when non-linguists try to use linguistic data to prove their pet theories. (See also Mark Liberman on the "file drawer effect".)