Moo und Bedeutung

« previous post | next post »

If you read German or Dutch, you may be interested in the recent spate of articles about the newly-compiled Cow/German Dictionary.  I'll wait to comment until after the BBC has scrutinized the story.

No, really, I'll wait until after today's meetings are over, and I've caught up with a few chores after taking the red-eye home last night from Albuquerque via Phoenix. Anyhow, I'm guessing that the dictionary's author, Gerhard Jahns, didn't get a press release into the channels that the BBC reprints…

In fact, Dr. Jahns' research seems to be a serious and long-established project that has reached a new stage, rather than the cheese-company PR stunt behind the BBC's previous cowlingual scoop. At an earlier stage, Jahns' work got extensive coverage back in 2002.


  1. Amy West said,

    June 2, 2010 @ 8:52 am

    I think you need to capitalize that thar "Bedeutung." Unless not capitalizing nouns is now the way things are done in German after the big spelling reforms, in which case I should get myself up to speed via the Kuh/Deutsch dictionary

  2. Carsten said,

    June 2, 2010 @ 9:03 am

    Amy: Capitalization is still done, at least in formal contexts.

  3. bulbul said,

    June 2, 2010 @ 10:34 am

    I'm assuming it comes from the authors of such lexicological masterpieces as Frau-Deutsch/Deutsch-Frau or Chef-Deutsch/Deutsch-Chef. However, I do have my doubts about their marketing strategy: Whereas the two named volumes have filled a serious void, the utility of their latest tome somehow escapes me.

  4. Marinus said,

    June 2, 2010 @ 10:55 am

    Since it's so short, here's my translation of the Dutch report found at

    The German biologist Gerhard Jahns has, through computer methods, compiled the first the first dictionary for understanding cows. This is another example of his years-long obsession with animal communication. Jahns hopes that his work will allow farmers to care for their cows better. The dictionary consists of about 20 words.

    Jahns performed a study involving 38 cows, recording their sounds with a microphone. These 700 recordings were fed into a modified speech recognition programme, which sorted them out into a series of 'words'.

    According to the German biologist, he has discovered, amongst others, the sound for "my udder is full, milk me" and "I am in hear, fetch a steer to cover me". The system that Jahns has developed appears to be accurate to 90%.

    You're not only dealing with my failings as a translator there, but also the fact that that report is written rather clunkily. I imagine the 90% accuracy is in matching the sounds the cows make to the condition the sounds are supposed to represent. This isn't a language by any standard, no more than road-signs are a language, so the press might (shock! horror!) be misrepresenting the findings a little (perhaps its the reporters, perhaps its the people handling Jahns' press). But I imagine it could prove quite valuable in the way Jahns intends: as a way to better read the conditions cows are in.

  5. Marinus said,

    June 2, 2010 @ 10:58 am

    Sorry, I see there's two typos in there. 'the first the first' is repetition, of course. Also, one of the delightful phrases that Jahns claims to have deciphered is: "I am in heat, fetch a steer to cover me". I find something Monty Python-esque of someone hearing a cow sound, looking it up in this cow dictionary, and finding that entry.

  6. Mark P said,

    June 2, 2010 @ 11:04 am

    I see no reason to doubt that cows would have several distinct vocalizations for specific conditions, but I wonder what they could have to talk about that would require 20 of them that are truly distinct and identifiable.

  7. Vicki Baker said,

    June 2, 2010 @ 11:08 am

    This is cool! When will a version of this technology be available for English-speaking cows? :-)

    One of the German articles says that Jahns hopes to work on a system that would pinpoint when a beehive has been infected by one of the parasites that cause colony collapse disorder. Awesome.

  8. Ray Girvan said,

    June 2, 2010 @ 11:14 am

    I rather imagined they'd speak lolcow: "I is in heat: I can haz bull cover me?"

  9. Vicki Baker said,

    June 2, 2010 @ 11:17 am

    The above info was from the "Die Welt" article, which I actually thought did a good job of explaining what Jahns was really up to – a system for monitoring animals cheaply and unobtrusively (to the animal) and that in creating this system he had "so to speak" created a "lexicon" for cows.

    But does the fact he still has a 10% failure rate mean that we can't ever really know what a moo is, and so can't measure the number of moos in Cow?

  10. John Cowan said,

    June 2, 2010 @ 11:28 am

    90% is pretty good for first-generation speech recognition, I'd say.

    [(myl) Or maybe not. If we don't know the details of the testing procedure, the number is nearly meaningless.]

  11. Chris said,

    June 2, 2010 @ 11:52 am

    I can' help but link to the 2009 Ignoble Prize Winner in the category: "VETERINARY MEDICINE PRIZE: Catherine Douglas and Peter Rowlinson of Newcastle University, Newcastle-Upon-Tyne, UK, for showing that cows who have names give more milk than cows that are nameless."

  12. Vicki Baker said,

    June 2, 2010 @ 12:16 pm

    John Cowan – I agree! I was making a facetious reference to the previous thread on word counting.

  13. Vicki Baker said,

    June 2, 2010 @ 12:18 pm

    Ray – when the farmer brings out the artificial insemination tube, do they say "Do not want!111!!!!"

  14. Josh said,

    June 2, 2010 @ 12:56 pm

    Some interesting claims made in the German article:

    Cows don't have vocalizations for pain
    He's been able to identify vocalizations for:
    Hunger/Thirst – 100% accuracy
    Sick and Coughing – 93% accuracy
    Milk/Udder problems – 74% accuracy

    The 20 moos he's identified are based on the the energy of a moo at a particular frequency (not 100% confident of my translation on this one)

    He's also looking to expand this to other animals. Pigs, bees, and Orca whales are mentioned in the article.

    He appears legit, and has been doing his research with government and university funding, but I didn't see any mention of peer-reviewed research. This could just be a cash-grab for his cow-moo-identifying system.

    [(myl) What Google Scholar knows of his published work seems to be most about tractor navigation, with a bit on the side about automating tomato quality grading. This suggests that the bovine vocalization classification might indeed be an entirely commercial enterprise, without any published discussion, peer-reviewed or otherwise.]

  15. AlexB said,

    June 2, 2010 @ 1:23 pm

    A puzzling difference between Dutch and German reporting is that the Dutch reports claim that Jahn has found 20 sounds, the German report mentions 10. Did something get garbled in the translation?

  16. anon said,

    June 2, 2010 @ 1:32 pm

    Also, one of the delightful phrases that Jahns claims to have deciphered is: "I am in heat, fetch a steer to cover me"

    Surely they'd be asking for a bull rather than a steer….

  17. Marinus said,

    June 2, 2010 @ 1:42 pm

    Shows you how agricultural I am: I didn't know 'steer' was closer to 'oxen' than 'bull'. In Dutch all male cattle are 'stier/en' — a nice false friend.

  18. Sili said,

    June 2, 2010 @ 2:47 pm

    According to the German biologist, he has discovered, amongst others, the sound for "my udder is full, milk me" and "I am in hear, fetch a steer to cover me". The system that Jahns has developed appears to be accurate to 90%.

    To the extent that 'binding barns' are still common, I can see the use of this. Most herds have grown too big for the farmer to learn these cues for individual cows.

    But free barns with automated milking are growing increasingly common. Likewise feeding can be monitored very closely and pedometers are know used to check the activity levels of individuals as well – cows in heat grow restless.

    But if this can be done cheaply and reliably, I guess it can possibly help to improve animal wellfare. I recall when the installation of an automated brush in a barn was enough to make the (local) news, but now almost all herds have them – they're cheap and make cows happier.

    But this is LanguageLog, not Dairists Daily. (Incidentally, if any readers do know of a online community of dairy and meat farmers, I'd be glad to hear of it.)

  19. Private Zydeco said,

    June 2, 2010 @ 5:41 pm

    Side note: "Mou" is "frown" in Cowish.

  20. Nanani said,

    June 2, 2010 @ 9:17 pm

    What does the system make of the Secret Cow Level in Diablo II, is what I'd truly like to know.

  21. Qkumber said,

    June 3, 2010 @ 8:51 am

    Has anyone else found themselves wondering at whom (and on what basis) the cows are directing these requests/pleas/bits of small talk?

  22. Bill Walderman said,

    June 3, 2010 @ 1:33 pm

    "Has anyone else found themselves wondering at whom (and on what basis) the cows are directing these requests/pleas/bits of small talk?"

    It's entirely plausible that the physical, acoustical characteristics of cow noises correlate with specific situations, such as full udders or estrus, and the research may (or may not) have confirmed this. That doesn't necessarily mean that anything like intentional "communication" directed at other beings is going on. But it's amusing to talk as if it were.

    [(myl) Exactly. An experienced mechanic can learn a lot by listening to the sounds that an automobile or truck makes; and no doubt this sort of acoustical diagnosis could be automated. This is a good example of what C.S. Peirce called an "indexical sign" (the canonical example is smoke as an index of fire), but it is obviously not an example of communicative intent.

    In The Expression of the Emotions in Man and Animals, Darwin observed that in the case of animals, such indexical signs can become independent of their original causes, because "Serviceable actions become habitual in association with certain states of the mind, and are performed whether or not of service in each particular case". This habitual association also need not involve any communicative intent, even if it gives rise to a communicative function that reinforces the association.]

  23. Jelle Zuidema said,

    June 4, 2010 @ 7:06 am

    For what it's worth, I wrote a story (in Dutch) for the Dutch weekly Intermediair on Jahns' Moo-classifiers back in 1997. It's depressing to find that popular science stories on "scientific breakthroughs" can easily be recycled 13 years later…

  24. Qkumber said,

    June 4, 2010 @ 8:09 am

    Thanks for pointing out the comic possibilities here Bill :-), but a serious thank you for an intriguing reply… If there is a connection between cow state and cow noise, would this be as a result of some evolutionary advantage? If that were the case, could something like that have been a step on the way to language? For humans, not cows! :-)

  25. Bill Walderman said,

    June 4, 2010 @ 4:25 pm

    "If there is a connection between cow state and cow noise, would this be as a result of some evolutionary advantage? If that were the case, could something like that have been a step on the way to language?"

    It's possible that correlations between cow states and cow noises could be the result of some evolutionary advantage in a pre-cow environment (cows are, after all, a human artefact). However, I think (and don't mistake me for an evolutionary biologist or someone who can speak with authority in that area) that the theory of natural selection doesn't necessarily demand that all features of animal physiology or behavior must be the result of evolutionary advantages–some features can arise by random processes and can persist as long as they aren't positively detrimental to individuals. As for human languages, I think that correlations between animal noises (or other types of animal behaviors) and specific situations are common throughout the animal kingdom, and in some species intentional communication may actually be involved. But the principal characteristic that distinguishes human language from animal vocalizations is syntax, and the origin of syntax is still a mystery that situation-specific animal vocalizations don't explain.

  26. Private Zydeco said,

    June 4, 2010 @ 4:55 pm

    "Moue", if one likes.

  27. Private Zydeco said,

    June 4, 2010 @ 5:45 pm

    Surely the bovine vocal apparatus did not evolve into being
    what it is — or rather, already was by the time animals of
    same genus and/or species were first domesticated by humans
    — without there being a survival-advantage of some kind in-
    herent in its use. One alternate theory being that it was mere-
    ly an adventitious event and not an adaptation in the strict
    sense — pardon the use of a "plebeian" mode of vernacular
    on and off here — but there must be -something- to it, right?

    McKenna states that human vocalisation has been known through
    sympathetic resonant frequencies to dislodge deposits of toxic
    metabolites and such from the cerebro-spinal fluid, and, in
    that way, it survives with yet another known function. Similar
    theories have been posited to account for the feline purr; it
    is a healing mechanism that uses sound waves to stimulate other
    body organs and tissues — vocal magic fingers, as it were.

    All that is somewhat a digression, but if it were only for this
    reason that cows evolved to moo and sheeps to bleat and so on,
    there would be that much in the phenomenon still. But to aver
    that cow-vocalizations have absolutely no semiotic meaning is
    going a bit far. Granted, they lack an extenxive vocabulary,
    and perhaps it IS mostly for catharsis — or, anyway, what ends
    up being catharsis — that any animal equipped for vocal action
    use what rudiments it owns. But lowing for a lost calf, or as
    one's hoof is being stepped on, or as one's nemesis approacheth,
    is at least a multi-role signalling mechanism if not a very
    accomplished system of signs.

    The car analogy is fitting, but also to a point. That is, Audi-
    torily assessing a malfunctioning automobile by engine noise, etc.
    would be, in terms of veterinary medicine, like listening to a
    heartbeat or to stomachic turbulence. As it is, just fixing the
    horn on the car is not fixing the whole car, or even a brake line.
    A car is not a psychologically autonomous entity in the way
    that a cow is. Feeble and inexpressive though that psyche may
    seem to humans to be, if a cow has been — through trial and
    error over the millenniae — programmed such that it does low
    when its hooves are stepped on, or as nemesis approacheth, or
    as its incapacitatingly overgrown mammaries approach rupture…
    they must mean something, Even if they don't know precisely how
    to say what they mean, or even what it is that they would mean
    if they could.

  28. Private Zydeco said,

    June 4, 2010 @ 5:54 pm

    Horns up!

RSS feed for comments on this post