Gerrit Storms reports on an interesting linguistic research project in which you can participate! Here’s the description:
Over the past few weeks, we have been trying to set up a scientific study that is important for many researchers interested in words, word meaning, semantics, and cognitive science in general. It is a huge word association project, in which people are asked to participate in a small task that doesn’t last longer than 5 minutes. Our goal is to build a global word association network that contains connections between about 40,000 words, the size of the lexicon of an average adult. Setting up such a network might learn us a lot about semantic memory, how it develops, and maybe also about how it can deteriorate (like in Alzheimer’s disease). Most people enjoy doing the task, but we need thousands of participants to succeed. Up till today, we found about 53,000 participants willing to do the little task, but we need more subjects. That is why we address you. Would it be possible to forward this call for participation to graduate and undergraduate students who are fluent in English?
The task can be found here.
I tried it myself—it only took a couple minutes. I only had two problems:
1. For each word, you are supposed to give three spontaneous associations. I find it difficult to give more than one, or at times two, spontaneous word associations. After that, it’s no longer spontaneous. I started to get tangled up in a concern of whether I should be giving synonyms or just related words.
2. My internet connection was slow when I was filling out the forms. Sometimes I was clicking and nothing was happening, other times it whipped through words too fast for me to follow.
P.S. Gerrit also writes:
If people would REALLY like to help us, they can forward the call to students, friends, family, etc. or distribute the call through facebook, twitter, etc. (In this way, we succeeded in building a word association network in Dutch over the past years. The network comprises about 13,000 words and was built using more than 4 million word associations, gathered from 100,000 native Dutch speakers. The problem is only: who cares about Dutch data. That is why we want to do the same in English.) Any suggestion about how to reach more participants is welcome (societies that we can e-mail, local communities who want to put this on their website, . . .)
Of course the network will be freely available to all interested language researchers when it becomes substantial enough.
Hi Andrew,
Funny title of the post. How did you chose this title? I am asking this because it is a common mistake in Dutch. The proper question is “Hoe heet jij?”. Strictly speaking is “Hoe noem jij?” the wrong question. However it is used so much that many aren’t aware that this is not correct.
The verb “Noemen” is used when you give a name to something, whereas “Heten” ask what name someone or something has (it has already been given). So when asking the name of someone, “heten” is the proper verb.
BTW: I remember you mentioning that you wrote an article in Dutch (with a little help?)
Louis,
Ja, I know that the use of noemen here is an unapproved colloquialism.
I think I only came up with three for one of them, and the others were about split between one and two. As soon as I thought about whether an answer was spontaneous or not, I decided it probably wasn’t. Also, I’m kind of surprised they’re not narrowing the geography down any further than they are. It’s only one more field on the start page, but a postal code or sub-region question could provide plenty of interesting additional data, I’d think.
At least as interesting would be seeing the variability of the associations of different individuals. I’ve read of some research that shows a great deal of non-overlap between people.
I completed the study but didn’t have trouble coming up with 3 words. I believe my problem will be how related they really are to others in my fuzzy brain. Many of my words that connected would be realted in some sentence or to my past experiences. Noteably, the word was increasing and I put function.. :)
I also found that most of my immediate associations were very influenced by temporary associations. Things like the podcast I listened to this morning, something I read last night, or a conversation i had been going over in my mind. I’m sure my answers would be quite different if I took the test a year, or even a day from now. It would be very interesting to expand the study by getting two or more responses from the same people over a period of time.
For me the associations tended to be compound words wherever they existed. e.g. “washing” led to “machine” and “powder”. “shuttle” led to “space” etc.
I wonder if they will be confused by my one of my pairs: dagger –> board. I have no idea why, but dagger made me think of daggerboard.
They did look at the stability of word associations over time.
There are probably more recent studies available, but in an impressive overview study Phebe Cramer concludes that in successive association, (i.e. with a period of several weeks) “up to 50% of the responses in normal subjects might chance. Importantly, the change is most likely to occur among those responses low in the hierarchy (so the less common ones) and at the same time, the overall response commonality increases” (p 23, Cramer, 1968).
I think it is hard to avoid common responses as ‘red’ to ‘rose’ etc., but I also agree that there is certainly a temporal effect that allows us to access a wide variety of weak links. In general, our results show that this added variability due to different contexts seems to give a better account for word processing and semantic tasks.
thanks for posting the call!
here are a couple of replies to what people commented.
1. we hired a server in the US to speed up the application. mostly it runs smoothly, but we’ve had some comments from users that it slows down sometimes. at some point, it was because we had as many as 450 users at the same time. we do check it regularly, though not constantly.
2. it’s not easy to say where spontaneous associations stop and an intentional search starts. some people mentioned that their second and/or third associations were sometimes second order associations (i.e., associations to associations). that phenomenon has been described, for instance by douglas nelson (USF), and we can see that it also occurs in our dataset. but we found that the second and third association does yield a richer semantic content than just the first. take, for example, blood. most people will respond with ‘red’. but the second response of most people is not a chained response (like ‘orange’ or ‘rose’), but rather things like ‘hurt’, ‘pain’, ‘wound’, etc. in a paper we’re writing right now we show that a network that takes into account the second and the third association yield better predictions of semantic tasks than just the first response.
anyway, thank you all for participating. this is a very exciting study for us and we’re glad that we got about 55,000 participants in the english study in just 3 months time.
if you enjoyed the task, don’t hesitate to forward the call to friends, families, students, and nice colleagues. if you hated the task, just forward the call to your enemies, … and to the other colleagues!
cheers,
gs
The thing the cynic in me worries about is they take 10 minutes from 150,000 people and I hope they get something useful out of it. The fact that this helps us understand Alzheimer’s is what, a leap of faith?
My first semantic memory of trying to learn a Germanic language such as Dutch is that the English “to learn” and “to teach” are often mixed up, thus the statement below from Gerrit Storms:
############
“Setting up such a network might learn us a lot about semantic memory, how it develops, and maybe also about how it can deteriorate (like in Alzheimer’s disease).”
############
My next association was with “to borrow” and “to lend” which also confuse Germanic speakers who are fluent in English. My third association is that “apparently” is rendered in Dutch opposite to its meaning in English. A fourth association is that the phrase “Dutch treat” has a totally different implication in America compared to the implication in the UK.
[...] Gelman posted our study on his amazing blog! Together with a number of interesting comments, it is definitely worth a [...]