,
12 tweets,
5 min read
Read on Twitter
Mini thread/rant about anthropocentrism in comparative communication: an increasing pet peeve of mine is our tendency to explore other species communication as it relates to ours through a human lens e.g. most studies are on vocal communication b/c of our speech bias 1/n
But we have to do this right, much easier to get that grant if we chuck 'evolution of language' in somewhere right? I don't think so. Studying species communication from the perspective of the species using it 1. allows us to explore awesome species specific capacities AND... 2/n
Critically, if we do it properly we'll understand what is meaningfully (rather than superficially) similar or different about human & other species. We can be anthropocentric about the question (if that's what rocks your world), but it's crucial that we're not in our methods 3/n
To illustrate it here's an example of where we messed up & are trying to do better #teamGesture We used to (like everyone else in the ape field) discriminate gesture 'types' (Arm raise, Arm shake...) on the basis of 'yeah.. that kinda makes sense to me' less than ideal right? 4/n
Many of us (fairly) independently landed on similar repertoires - so it seemed to make sense to everyone. But everyone making those choices was a human. What if we were splitting/lumping on the basis of something totally irrelevant to the other apes using these gestures? 5/n
Imagine not knowing if tone meant change in emphasis (eg English/German) or change in meaning (eg Thai/Vietnamese) - it'd be pretty hard to do any meaningful analysis of the signals if we can't parse them into relevant chunks [this used to legit keep me up at night] 6/n
Bit of a mess to be in after already working on gesture for 10+yrs 😱 but... we've been working on this for 10+yrs - we don't need to use our behaviour to parse anymore we have 1000s of cases of the ape's behaviour 🙌 7/n
We used the descriptors from our repertoire (movement, body part, repetition, ...) to create a theoretical maximum possible repertoire (all of the possible combinations multiplied) -> that gave us >6000 gesture types 8/n
But..! Many of these were physically impossible (e.g. movement = spin + body part = head). If we ditch these we only have just over 1000. Of which we actually only see ~124 in the apes [side note: apes use only a fraction of the ways they can move their bodies as gestures] 9/n
We looked at whether there was a reason to lump these 'types' because the apes' behaviour suggested that they were treating them as the same signal. Eg does it make a difference if you SHAKE + ARM or SHAKE + HAND? Pretty graded right, should be the same? Nope, not to a chimp 10/n 
We ended up totally reclassifying our repertoire, including more than a few gestures we used to lump & splitting new ones. And b/c we have less data for some gestures than others it's still a work in progress (paper tinyurl.com/y2c2z9as OA tinyurl.com/y69ca6nf) 11/n
But... we finally have a repertoire that makes more sense to the species that's using it AND we can make more meaningful comparisons w/ human language as a result (b/c I guess that's kind of fun too..🤘) #winwin #teamGesture greatapedictionary.com 12/fin
