Hacker News new | past | comments | ask | show | jobs | submit login
Experimental evidence for compositional syntax in bird calls (2016) (nih.gov)
142 points by studius 4 days ago | hide | past | favorite | 40 comments





I did a hike on a mountain ridge in Hawaii (Stairway to Heaven), and was very shocked by how loud the highway below was. I was maybe 3000 feet above. The sound was amplified from from the valley below, the mountain sides almost acting like walls of a speaker.

It's something that I don't think gets enough attention, the way light pollution does. But its' definitely a problem in some places. And I can see how it would disrupt bird calls.


It's something that I don't think gets enough attention, the way light pollution does.

It does in German speaking parts of Europe - highways on weekends and during the night have speed limits set to reduce noise.


I think that is because of the people living near them. Were the birds even considered?

So true. I am rather shortsighted, but my body compensated by developing quite good hearing sense. The noise pollution all around is serious.

That was pretty noticeable during the lockdowns earlier this year. It was really quiet.

Humans have tonal components to meaning, e.g. in English, "You are hungry." doesn't have the lift at the end as does "You are hungry?". I wonder if any birds do something similar to distinguish, e.g. "Predator!" vs "Predator?". Maybe someday an app on our phones will be able to translate as we walk through the woods, like "A female Carolina chickadee in the sycamore at 10 o'clock says there's a raven coming from the south".

There are features in the frequency ranges of bird calls that suggest this. Alarm calls are usually on frequencies that have better propagation (this also depends on the habitat that birds occupy - ground living birds have much lower frequency calls because the sound spreads better close to the ground. Calls that go to chicks tend to be on higher, less propagating frequencies. On the other side, the calls that hunting birds make tend to look like an all frequency sweep - possibly they are trying to get a response from other species to locate them.

This is the case even though English isn't even considered a tonal language. Mandarin is a highly tonal language with four distinct formalized tones. Not only that, but the brain has to adapt to pick out those tones clearly. As a result, "perfect pitch" is much more common for Mandarin speakers than speakers of non-tonal languages.

Tonal here is very different. The tone in Chinese is a completely different word. Multiple words have the same tone and thus pronunciation with different meaning (and character). Tone usage on one word would not change the interpretation.

There are cases where this is kind of like English, in the case of buy/sell 买/卖, but still quite different


Yes, that's why I referred to Mandarin as a "tonal language". It is a formal linguistic categorization. Speakers of all languages may use tone, but an actual tonal language is quite different, taking the concept of tone to a another level in terms of grammatic and lexical importance.

> "Predator!" vs "Predator?"

Should have gone with "Friendship!" VS "Friendship?"


There really are birds that have predator warning calls. Other animals even listen for them and hide when they hear them.

We can always tell when our neighbor's cat is outside when the birds in our yards start to get agitated.

I've noticed the warning call differs at times, I hear two chirps for land based predator and three for air. Small sample size, city finches.

BNF grammar and sample inputs/outputs: https://paste.debian.net/plain/1167596

If your parser can't handle that, then it's literally dumber than a tit.


I may be misunderstanding, but didn't the paper say the Warning has to come before the Approach to elicit much of a response?

Yes, ABC-D elicits a response while D-ABC doesn’t

Just curious, are there any projects for creating a software talking parrot? Maybe one that would integrate with Alexa of Google voice assistant.

Video is 5 years old. Seem pretty cool. Steampunk parrot. https://www.youtube.com/watch?v=6mxSl06IZaY

Edit: I guess this is the guys original channel https://www.youtube.com/watch?v=esSbc4phCjk&t=34s


TLDR: they have studied two types of calls in a specific bird - one type is dubbed ABC and elicits a horizontal scanning behavior; the other, dubbed D, elicits approaching the loudspeaker. Then, they have shown that the sequence ABC - D elicits both behaviors at the same time (the birds scan the area and they come close to the speaker). However, the sequence D - ABC does not elicit this combined behavior - it only elicits one behavior or the other, but not both in the same individual.

The conclusion is that the birds interpret the sequence ABC - D as a meaningful unit, different from its constituent parts, which is a kind of interpretation only previously confirmed in humans.


We have a parrot and he definitely understands the difference between “put kiwi to bed” and using “bed” in any other context.

Of course it could also be because he correlates environment cues to bed time.

Either way, birds are smarter than you think until you spend time observing them.


  > birds are smarter than you think until you spend time observing them
People, too.

I've been wanting to get a parrot for a while. How often do parrots end up speaking like colorful, foul-mouthed sailors [1]? Seems charming, honestly.

[1] https://www.mlive.com/news/2020/09/zoo-separates-foul-mouthe...


Depends on the species. Mine doesn’t talk, he instead learned to whistle. His species (sun conure) doesn’t normally whistle, he picked it up from me. It’s more pleasant than the 80db screams so we encourage it.

He used to say his name sort of (if you added some human pattern recognition magic) but stopped after we moved apartments. No idea why.

The worst is when he mimics human speech patterns. You can tell it isn’t words but it’s close enough to distract you every time.

Main thing with parrots is that they’re more work than people think. Like human children they get bored, have tantrums, internal emotional states, and respond to the same stimuli differently based on mood. And unlike children they cannot be reasoned with. Some are smart enough to lie and obscure because they have theory of mind.

And they stay like that for 30 years. Some species even longer.


I'd suggest you do some 'parrot sitting' first, look after someone's parrot for a while to see if it is something that you will really like. A lot of parrots get bought as impulse buys by people who have absolutely no idea what it is like to actually have a parrot and to have to take care of it.

Yeah, that's always the right answer. I've gone through this mental exercise multiple times before -- I want to get a parrot! But they live a long time and are extremely intelligent. They're very independent and when they are unhappy can become destructive. Am I a good enough beast-keeper to pull that off? Not really. I agree with you that parrot sitting is a smart idea.

I wonder how generalizable research on parrot language would be to other bird species? They seem to be pretty unique in their capabilities, but perhaps that's just from a human perspective.

Some corvids can learn how to speak but I haven’t heard whether they understand the words. With parrots there’s evidence that some species learn meanings of the words they use.

Even the cortical neurons of ferrets have special representation for phonemes from human language: https://pubmed.ncbi.nlm.nih.gov/18247893/

Every time I read one of these articles, I start feeling increasingly worse about not being a vegan.

I still eat meat, but I do wonder what paper is going to get released that is the tipping point for me; it's easy for me to justify eating meat if I think these animals are so stupid that they can't feel higher-level emotions, but if we have legitimate evidence of animals having "uniquely human" traits, I can't help but think that us eating them is sort of ungodly cruel.


I empathize, but I'm also curious -- does it change anything for you that many bird species are themselves carnivorous?

It is a factor I think about; I'm not saying that it's never ethical to eat meat; I don't condemn a lion for eating an antelope, since I'm not even sure it's possible for a lion to be a vegan if it really wanted to [0]. Similarly, I don't have a problem with birds being carnivorous.

That said, anyone reading this post is probably not a lion or a bird, and has some level of self-reflection as a result. Just to play devil's advocate for a bit, couldn't you argue that if it's possible for me to live a fairly healthy life without killing any higher-minded creatures, then I should? If we are smart enough to have an entire philosophy of ethics, then shouldn't we be using this ethical framework to minimize suffering in the world?

I should reemphasize, I'm not a vegan, or a vegetarian, I have a fairly typical American diet of "it's not a meal without meat" sadly. I'm just saying that these kinds of papers make me question that.

[0] I'm not a biologist or doctor or anything, but my understanding is that cats (at least house cats) cannot product their own vitamin A out of Beta-Carotene like humans can, and must get it out of animals. I certainly cannot blame any animal that is doing what it absolutely has to for survival.


Thanks for the thoughtful response! To be clear, that wasn't a loaded question -- as someone who hasn't thought too hard about this, I really did want to hear your opinion. :)

I believe humans still require vitamin B12 which we can only get from meat, however you can supplement this via plant-based milk and other ingredients.

B-12 is most readily available in meat, but is bio-available to reduced/smaller levels from non-meat sources (e.g. some fungi and algae).

Obligatory reference to Raymond Smullyan's seminal To Mock A Mockingbird [1], one of my favorite books and a great introduction to the lambda calculus. I guess the resemblance to actual bird language is more than just passing.

[1] https://en.wikipedia.org/wiki/To_Mock_a_Mockingbird


Yes, I just thought about that one. That's funny.

Am I the only one wondering what the National Institutes of Health has to do with ornithological research? Isn't this a little outside their mandated field?

The NIH did not run this study. The submitted link is just the article (from the Nature Community journal) being hosted by the NIH's library.

https://www.ncbi.nlm.nih.gov/pmc/about/intro/


This work was done with a combination of funding from the US national science foundation and Japan's JSPS (see acknowledgements) and doesn't involve the NIH directly. PubMed/PMC are hosted there, though.

That said, animal models are extremely common for NIH-funded work, and birds feature prominently in studies of language and auditory processing.


I don't know the details of this study, but aside from that this might prove relevant to humans, NIH also do veterinary research: https://www.ors.od.nih.gov/sr/dvr/Pages/default.aspx



Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: