restructr!

mikroblogolas:

Warning: Linguistics ahead

You’ve probably seen the above video. It’s called “twin baby boys have a conversation” and has over 13 million views so far.

Geoffrey Pullum of Linguist Log notes, “It isn’t as funny as the Ranting Toddler, but it sure does suggest that the idea of conversation, complete with intonation and hand gestures, emerges way before even a single word is learned.”

As I watched the video, I started wondering about the sounds they’re producing. They’re repeating more or less the same syllable the whole time. At first it sounded like da da da. The vowel is uninteresting enough; it varies a bit but is usually around a mid/central vowel. Not hard to produce. But d? That’s a voiced stop. And languages tend not to use voiced stops if they don’t have to (despite the fact most Indo-European languages, including those of power and influence, do happen to have a voicing contrast). Voicing is an effort you don’t need to go to if you don’t want to make a distinction. So why would babies, obviously not speaking real English, clearly not forming voiced and voiceless minimal pairs, bother with voicing?

Then I kept listening: Maybe it’s a voiceless stop. An unaspirated t. It definitely sounded like it could be. (The sound certainly isn’t aspirated.) Why hadn’t I heard that before? I started thinking about how unbelievable it is that I as an English speaker could be so bad at hearing the difference between an unaspirated t and an English d. It’s not that I don’t speak languages that have a purely voicing distinction, because I do (Hungarian, Spanish). But somehow, English, with its not really voiced d’s and b’s and g’s, makes me hear those semi-voiced sounds instead of unaspirated t’s and p’s and k’s.

(Background on the voicing contrast for stops in English: In the minimal pair pie~buy, the difference is not only voicing but also aspiration. The voiceless stop (here p) is aspirated, while the voiced stop is not. This means that the voicing distinction does not have to be as strong. And in fact, English voicing is delayed and thus weaker than in languages in which the only distinction is voicing (e.g., Spanish: cata~cada) and the voicing time is longer.)

I had watched the video assuming the babies were English speakers. The babbles sound like they could well enough be English-like baby talk. If had been expecting another language, would I have heard the voiceless t? Would speakers of other languages hear a t?

And then I checked the comments. (The video, by the way, has been seen all over the world. Check out the demographics map. I guess it’s because it “speaks” to everyone.)

One YouTuber transcribed the video like this in his comment:

ta ta tatata ta tat atataa??? tata tat atatat tatatataaaaaaaaaaaaaaaaaaaaa!!

A voiceless t! Can’t be a native English speaker. I checked his profile, and he’s from Indonesia. (Indonesian, not surprisingly, doesn’t use aspiration in its stops.)

Another commenter, this time from the Philippines, transcribed it with a t as well:

TATATATATATATATATATA? TATATATA TATATATATATATATATATTA!!!!!!!!!

­!!!!!!!!!!

But surely others heard it, like I first did, as a d.

One such commenter wrote:

Conversation before the camera started recording….

Baby 1: Please, can I have a raise?

Baby 2: Hahahaha, nooooo you can’t…

Baby 1: Give me a raise!!!

Baby 2: Why?

Baby 1: Ehm….. Because I am crazy, dadadadadadada!!

Baby 2: Hahaha, you call that being crazy? This is crazy, dadadadadadada!!

I checked to see where this commenter is from: Denmark. (I don’t want to assume native languages from a person’s geographic location, but from the profile, it seems that this person does speak Danish. Likewise for the commenter from Indonesia.) And guess what: Like English, Danish uses aspiration for the voiceless stops and no aspiration for the voiced stops.

A couple more people from English-speaking places (United States, United Kingdom) wrote it as da da da. (At least two commenters on the Linguist Log post, presumably native English speakers, also described the sound as da. No one in the comments suggested the sound could be a ta.)

It would be interesting to investigate this further. How do different people perceive the sounds the babies make? How is this influenced by the languages they speak? How is this influenced by the language they expect to hear? How “native” do speakers have to be for the phonologies of the languages they speak to affect their perception? Acoustically, which sound are the babies’ babbles closest to: an English d, a fully voiced d, or an unaspirated t? Is it dental or alveolar? Why, in terms of articulation, would babies choose to make this sound (over, say, labials)?

Time for YouTube phonology research! Unfortunately, it’s actually time for me to do real homework, which is, sadly, less exciting.