I’m guessing that by now most of you have heard about the Endress et al study that monkeys are able to learn certain mechanisms that underlie human language. I want to talk about that research, but I also want to talk about the way it’s being covered on BBC News. Last week, I saw a story whose headline read France ‘banned Yemen crash plane’; I misread the headline as France ‘banned Yemen plane crash‘, because I saw the quotes and figured that no one would ever say or write “banned Yemen crash plane” outside of a British headline. As often noted by Geoff Pullum, the British news uses this distinctive noun-compounding scheme in their headlines. Yemen crash plane is a stack of three nouns, referring to a Yemeni plane that crashed. It’s a somewhat reasonable means of compressing a lot of information into a headline, but it’s not how people talk, so I was surprised that anyone would be quoted saying that. And, sure enough, they weren’t. The quote was actually
“A few years ago, we banned this plane from national territory because we believed it presented a certain number of irregularities in its technical equipment,” Mr Bussereau told parliament.
Eh, it’s not a direct quote, but it’s the same basic idea. I don’t feel comfortable with this myself; I was taught in my high school journalism classes to believe in a Sanctity of the Quotational on par with that of the Confessional. I checked with a journalist friend at CNN and she agreed that what goes inside of quotation marks ought to be a direct quote, even in a headline. All in all, it’s not a big deal; the meaning of the headline quote and the actual quote are essentially the same, so no harm, no foul, I guess.
But yesterday I read the BBC’s take on Endress et al, the research that showed that cotton-top tamarins can learn to recognize a simple affixation rule. First off, let me note that this is a pretty cool result, and the paper is pretty easy to read if you want the details straight from the horse’s mouth. The researchers played recordings of nonsense words to a group of tamarins, where each word consisted of a variable stem and a constant affix. The tamarins were familiarized with a series of stems+suffix (e.g., bi-shoy, lo-shoy) or a series of prefix+stems (e.g., shoy-bi, shoy-lo). This familiarization was done without any incentive to learn; they weren’t given food or shocked or tested in any way. The words were just played over a loudspeaker in their enclosure. After the tamarins were familiarized with the words, they were taken to a new enclosure, where examples of suffixed and prefixed words were played to them (the affix remained the same as in the training period, but the stems were new). The tamarins stared surprisedly when shoy was used as a prefix when they were trained with it as a suffix. Similarly, if they were familiarized with shoy as a prefix, they reacted to its use as a suffix. In summary, their “results suggest that, in the absence of training, cotton-top tamarins learn a rule that is formally similar to affixation patterns […] in natural language.” This is a nice result in my book because it shows that
“[…] the language faculty uses similar positional mechanisms to compute affixation patterns, and though these mechanisms are uniquely used in humans to create and understand words, the mechanisms themselves are not specific to humans or language.”
Which is to say that some of the core pattern recognition systems that underlie our ability to learn language are not unique to humans. (I think this result makes it a bit easier to understand how language could emerge through evolution.) There’s a bit more that can be said about that point, but that’s the basic finding of the research. The researchers do a great job of establishing that this as far as the data should take us:
“[…] nonhuman animals may have the capacity to learn surface transformations involved in affixation, but they cannot link them to other aspects of linguistic structure. […] Unlike other primates, however, infants can use such evolutionarily ancient abilities for purposes that are specifically linguistic and (presumably) unique to humans.”
Which is to say that all the news stories talking about how monkeys can recognize bad grammar have it all wrong. Monkeys can recognize the difference between established and new morphological patterns, but there is no evidence that monkeys can expand this to a grammatical system. It’s right there in the paper.
The BBC, though, uses the headline Monkeys recognise ‘bad grammar’ . Why on earth are there quotation marks around bad grammar? Unlike the plane crash example, there is no quote in the story about grammar at all. The pre-print of the paper I posted to above also doesn’t mention grammar anywhere. (It does use the phrases “grammatical competence” and “grammatical computations”, but establishes immediately within these sentences that only the abilities of humans are being referred to.) Judging by how scrupulously the authors avoided referring to grammar, and how they restricted their conclusions, I would be surprised to find that they suggested in interviews that monkeys know anything of “bad grammar”. So why is bad grammar in quotation marks? Who is it supposed to be quoting? What use are the quotation marks if nothing in the story tells us who the quoted matter is attributed to? It’s misleading and inaccurate, and contradicts the findings of the research. And, of course, it’s the misleading headline that’s stuck in everyone’s minds. Alas.