Vocalizations of wolves and justices

« previous post | next post »

Tessa Koumoundouros, "Adorable Study Tests How Dogs Respond to Wild Wolf Calls – And, Yes, There's Footage", ScienceAlert 2/12/2023:

Without convenient access to phones for pens for letter-writing, wolves must rely on howls to communicate long distances. These woeful wails allow the social mammals to maintain their territories as well as keep track of and stay in synchrony with other pack members. […]

A new study exposes family dogs to wolf howls to better understand why some of our canine companions no longer seem to bother with this seemingly important form of dog communication.

The original press release, from the Institute of Biology at Eötvös Loránd University, is here. The scientific publication is Fanni Lehoczki et al., "Genetic distance from wolves affects family dogs’ reactions towards howls", Nature Communications Biology 2/6/2023:

This reminds me of a sad failure in my past, which started with a happy success.

About 20 years ago, Jerry Goldman at oyez.org was advised by someone at NSF to recruit me to create a program to do speaker identification on U.S. Supreme Court audio recordings. The reason, as explained in a paper where we described the (successful) results of our part of the project (Jiahong Yuan and Mark Liberman, "Speaker Identification on the SCOTUS Corpus", Acoustics 2008):

The U.S. Supreme Court began recording its oral arguments in the early 1950s, and some 9,000 hours of such recording are stored in the National Archives. The transcripts do not identify the speaking turns of individual Justices, referring to them all as “The Court”. Therefore, as part of a project to make this material available online in aligned digital form, we have developed techniques for identifying speakers and aligning entire (hour-long) transcripts with the digitized audio.

The successful results of the overall Oyez SCOTUS database creation are sketched in a later post "NPR: oyez.org finishes Supreme Court oral arguments project", 4/13/2013.

While we were working on the SCOTUS speaker ID problem, we were contacted by some folks associated with the project to restore wolves to Yellowstone National Park, who hoped that we would be able to create a "howler ID" program to identify wolves on the basis of their vocalizations. This would be combined with a mile-square grid of satellite-connected, solar-powered microphones, placed thoughout the park. The whole system would enable identified howls to be localized in space and time, thus replacing the system of radio collars then in use.

Why do this? As I understand it, the idea was to replace radio collars. To deal with concerns about wolves from the park killing livestock outside the park, each animal had been fitted with a radio collar for monitoring its location. There were two problems: batteries run out and have to be periodically replaced, which required the animal in question to be captured; and radio collars are not heritable, so each new generation had to be captured and collared. The animals were captured by shooting them with tranquilizer darts from a helicopter, which was expensive for the shooter and traumatic for the shootee.

A "howler ID" system should definitely be possible, since the social functions of wolf howls depend on the ability of other wolves to identify the howler.


  1. To get the idea off the ground, we would need to prove that computational "howler ID" really works;
  2. You'd need a reliable way to register individual wolves with the system.

Solving problem #1 would require a large enough set of howls from a large enough set of wolves. Luckily, an ethologist friend told me about a wolf study site, on an island in the St. Lawrence river, where it should be possible to record howls from known individuals. And the folks running the site were happy for us to come make recordings. But it would take a couple of months to get enough recordings, so someone would have to spend a summer on the island. For various reasons, I couldn't do it.

We tried to recruit a grad student, and failed.

This made me sad. That was partly because the wolf ID project would have been fun and interesting. But also, I was looking forward to describing my current research project as vocalization identification for Supreme Court justices and wolves.



  1. Taylor, Philip said,

    February 17, 2023 @ 3:59 pm

    "to phones for pens for letter-writing" — should not that first "for" read simply "or" ?

    [(myl) Probably, but your comment belongs on the ScienceAlert site, since the typo (if it is one) is theirs…]

  2. mg said,

    February 17, 2023 @ 5:03 pm

    Too bad. You could have had a paper titled "Rulings with wolves"

  3. AntC said,

    February 17, 2023 @ 11:07 pm

    spend a summer on the island. …

    We tried to recruit a grad student, and failed.

    Sheesh! I thought grad students were always complaining they couldn't get into research. Free Summer holiday, with nothing to do but roam about on a forested island; and only the mild inconvenience of possibly being torn to shreds by wolves. I'd have jumped at the chance …

  4. JPL said,

    February 17, 2023 @ 11:47 pm

    WRT the quotation from Science Alert, and having lived in a place where domestic dogs almost all are outside at night (when it's otherwise quiet), it's puzzling to read that, "… some of our canine companions no longer seem to bother with this seemingly important form of dog communication", when my experience was that they still do that. And how are they so confident about the dogs' reason for these vocalizations? Could it not just as well be that they are not intending primarily to communicate something, but are simply expressing something (the humans are interpreting them as expressing a "woefulness", but it's probably not really that), and that the communication is inadvertent, and that the other dogs are howling in agreement. (It's unlikely that the dogs I used to hear howling were familiar with each other, i.e., they probably couldn't identify the howlers and had never met them. "Territory" also does not seem relevant in these cases) Maybe there's an element of wonder at the big bright orb in the night sky, if that is the context. For us it's sunsets; for dogs it's the moon.

  5. Taylor, Philip said,

    February 18, 2023 @ 9:09 am

    "[(myl) Probably, but your comment belongs on the ScienceAlert site, since the typo (if it is one) is theirs…]"

    Ah, well, there was no "[sic]" inserted in the quotation, so I assumed that the typo (if such) had received editorial approval …

  6. Mark P said,

    February 18, 2023 @ 9:46 am

    JPL — Dogs bark for a variety of reasons. One of my dogs barks at deer, possum, and possibly ghosts in our front yard. She barks immediately when she goes out the door to the driveway. I think that’s reconnaissance by fire; she’s just trying to see whether something is outside and will respond. But she also definitley barks in response to other dogs’ barks in the distance. It’s a back and forth. I don’t know what is communicated other than “Hey!”, but it’s a communication of some sort.

  7. Seth said,

    February 19, 2023 @ 9:15 am

    Wouldn't it have been enough "proof of concept" do just do "bark ID" from known dogs, or "meow ID" from known cats? If you could either one of those, or better, both, that would seem to make it almost certain that "howl ID" for wolves was possible. And given the advances in computation power, and all the videos available online of dogs and cats, I would presume these are solved problems by now.

  8. Tim Leonard said,

    February 19, 2023 @ 11:14 am

    A lost opportunity to prevent some wolves from becoming Supreme Court justices.

    (Sorry; couldn't resist.)

  9. Andy Horn said,

    February 19, 2023 @ 3:06 pm

    For wolves, see Root-Gutteridge et all 2014. Identifying individual wild Eastern grey wolves (Canis lupus lycaon) using fundamental frequency and amplitude of howls. Bioacoustics, 23(1), pp.55-66.

    For the potential of auto i.d. in monitoring for mammals in general, see Linhart et al 2022. The potential for acoustic individual identification in mammals. Mammalian Biology, pp.1-17.

RSS feed for comments on this post