HN Explorer

18 comments

[ 2.1 ms ] story [ 49.8 ms ] thread
peterburkimsher
Let's build a bird modem, and let them join Twitter!

It would also be quite fun to make a user interface where a bird could "fly" through Google Street View, and sense the location changing based on an electromagnet instead of the magnetic north pole that they normally use for navigation.

Since August 2021, I've been trying to apply UncleBob's advice to "clean as you go" (Scout Rule) and picking up litter while biking to work.

https://www.youtube.com/watch?v=BSaAMQVq01E&t=2021s

The birds have been particularly friendly and thankful. Pigeons and Pūkekos help pick up food from the ground. They're all surprisingly trusting (especially if I have some fish & chips to share). Seagulls look down, and shout about litter. Sparrows are smaller, faster, and have lower energy requirements, so eat smaller pieces of food.

So while I'm going around looking to pick up biohazard waste (used masks in bushes), the birds are helping, and it would be so nice to let them chat to each other, even if we don't understand their QPSK or other encoding schemes. Parrots get so bored in cages, let's put them on the Internet :)

ChaitanyaSai
The song of plain-tailed wrens is fascinating. Male and female birds perform a high-tempo duet where it's hard to tell it is two birds doing the singing. https://www.youtube.com/watch?v=T3GprtBhekA

(We discuss songbirds and the importance of sequence building/remembering for language in our recent book. Here's a review: https://www.kirkusreviews.com/book-reviews/ogi-ogas/journey-... )

circlefavshape
Any progress in establish whether the song is, in fact, language?
ChaitanyaSai
ML techniques are leading to some interesting work in determining this. Here's one such example https://journals.plos.org/ploscompbiol/article?id=10.1371/jo... Female zebra finches do not sing, only males do. I'd imagine that if bird song is high on communication, it would be seen in both sexes (and there are song birds where females sing, but not studied as much as far as I know.) tldr: still unclear, but new techniques give us hope we can answer
ericbarrett
I wouldn't be so quick to write off birdsong as speech based on this one species—given the dramatic variability in form, behavior, and habitat across the animal kingdom, a particular family of birds where the males don't "talk" wouldn't surprise me at all.
ChaitanyaSai
Absolutely. Agree there's a lot of variability. Only pointing out that in species where only males or only females croon, it may be unlikely there is broadspectrum communication happening.
xoserr
I remember walking early morning last summer, listening to the birds and thinking how absurd it is to believe this isn't a type of language.

I guess it would depend on what is considered the minimum constraints to be considered a language.

totetsu
it's always fun to take a look at bird song recordings with Audacity spectrum view.
servytor
I remember part of this book called "The Symbolic Species: The Co-evolution of Language and the Brain", and the author got so angry when layman compared birdsong to speech. It was a very interesting book though.
jcims
More of an artistic exploration but I think somehow still gets at a similar intuition in this context. Oona Räisänen (aka windytan) did a small write up and some code about how human speech limited to a single harmonic actually sounds a bit like birdsong.

https://www.windytan.com/2021/03/speech-to-birdsong-conversi...

YouTube video at the bottom provides an example. She also makes code available to duplicate.

Samuslav
I live with a Green Aracari, and I swear there's more than just signal in the noise. They make fast-paced series of rattles and chattering sounds, where different patterns of pitch have different meanings-- reminds me of the "tonemes" of tonal human language. But there are without a doubt sounds that map cleanly to certain meanings- one sound in particular means hawk or "sky problem" and I can replicate it and get the bird into lookout mode immediately. The species is known for cooperative hunting (nest raiding), so it wouldn't surprise me if they had at least some basic communication patterns approaching language.
sdenton4
By the way, we're running a Kaggle birdsong species id competition right now, for anyone who would like to try their hand:

https://www.kaggle.com/c/birdclef-2022

The focus this year is on few-shot learning for rare+endangered species classification.

thro1
But.. in my garden (or parks) the birds, which sing together, every year have different song and kind of imitate/answer each to other, repeating it, progresively trough months: two years ago they started really lame with some stupid phrase like.. someones doorbell and made it better with time, last year I enjoy the melody since the beginning. If it could be an evolving story, the words which stayed were created ad hoc, for invisible, recorded reasons - if in a competition, different birds complete the song in own (but sometimes similar) way. If it's a language, its the same kind as the music would have ( in the sense https://www.classicfm.com/music-news/study-proves-music-is-u... ;). Other thing: the song is not always the same - so it may communicate something.