Slashdot Mirror
Dog Trained on 200-Word Vocabulary
An anonymous reader writes "The Max Planck Institute for Evolutionary Anthropology in Leipzig, Germany is reporting in Science Magazine today on an example of successful human to non-human communication: Rico, a collie trained on a vocabulary of 200 words. Their conclusion is that 'brain structures that support this kind of learning are not unique to humans...[Rico has a] retrieval rate comparable to the performance of three-year-old toddlers'. In case you ever wondered if your dog understands what you are saying, Rico 'can learn the names of unfamiliar toys after just one exposure to the new word-toy combination.'"
I just saw something on Deutsche Welle (in Los Angeles actually) and that dog indeed picked out a bunch of items among dozens littered across the floor on verbal request. What's interesting is that the canine still used his nose (not his eyes) to identify the object. Looks like his brain is correlating verbal commands with smells - contrary to how human beings would solve this problem. ;-)
Anyway, I never bought into that whole 'humans are unique' bullcrap - countless reports have proven that several species elicit signs of abstract thinking, verbal communication (whales, dolphins in particular), emotions like sadness (chimpanzees and other primates), anger, tendency for rape (chimpanzees again - why am I not surprised? LOL), etc.. Why are we still so full of ourselves and continue to describe ourselves as the crown of evolution while we decimate other species and commit atrocities unknown to any other species on this planet. I hope this dog doesn't smarten up too much - once he realizes how screwed up his 'masters' are - he's probably reconsidering that whole loyalty issue
Many german research centers are named after Max Planck. Google for "Max Planck institute" to find many many other fields Planck didn't do work on.
i should also point out that border collies are smart because they aren't standardized by the akc. once a "perfect" version of a dog is picked by the akc, they're inbred to keep the same look and they get stupider and stupider. there is an appropriate simpsons quote about the inbreeding of dogs (specifically, dalmations) but i can't seem to find it..
--BlueLines "The cost of living hasn't affected it's popularity." -anonymous
Last I heard the average human had a vocab of around 2500 words or less.
From an article that I read on this exact topic (the dog that is) a few days ago it claimed that the average high-schooler graduating has a 60,000 word vocabulary. A quick search on news.google.com found:
But Lori Markson of the University of California at Berkeley stressed that children develop a diverse and extensive language base. A 5-year-old child knows 7,000 to 8,000 words and what they represent. An average adult knows 60,000 words. Educated adults may know upwards of 100,000 words. Most of these words are learned after a single exposure, said Markson, who collaborated with Bloom on a study of fast-mapping in children.
I should have been more specific, 2500 often used vocabulary. I read it when researching Pimsleur:g erman.asp
http://www.pimsleurapproach.com/learn-
I'm trying to learn German so thats what I linked to.
Well, I posted essentially the same thought as you then - OK, after the fact - gave google a whirl and the estimates actually vary widely.
Fundamentally, I think it's pretty difficult to measure a person's vocabulary. Do we measure the range of words they use every day, or the range of words they might ever use, the words they understand out of all dictionary words, the words they kind of understand in context but couldn't give a definition for... and so on.
I think 5,000 might be reasonable for a daily-use vocabulary, and 25,000 sounds good for the number of words for which an individual can give a fair definition, 60,000 might be "rough comprehension".
For instance, many people will use the word "laminate" without being able to define the process of lamination. They might simply see it as the act of making a piece of paper shiny. Certainly, that's how I see it! And so on.
Yours Sincerely, Michael.
A lager IS a pilsener and vice versa.
This is bullshit, according to Geoffrey Pullum, professor of linguistics at the University of California, Santa Cruz.
Screw paraphrasing: " The trained object-fetching behavior of Rico, the border collie that this German research is talking about, has nothing at all to do with understanding language. The behavior is comparable to what you would have shown if you demonstrated that you had trained your goldfish to swim to a given object in its tank when you showed it a card with a given letter of the Greek alphabet. By all means attempt that too, if you think it would be interesting science. But don't bring it to me for my approval under a headline saying Research Shows Goldfish Can Read Greek, that's all! Unless you actually enjoy seeing the veins standing out in my neck as I hurl some more defenseless chairs and coffee tables and goldfish tanks around the room. "
His post is available here. And for those geeks interested in language, check out the Language Log.
Wow, so much ignorance in one post, where to start...
Border collies are AKC's 139th breed. Because it is a relatively new AKC recognized breed, "AKC will accept accept dogs registered with the American Border Collie Association (ABC), the American International Border Collie (AIBC), and the North American Sheepdog Society (NASD)." (AKC web site) It is currently in open registration (see web site for details-but requires the dog to have a pedigree, submit pictures etc.).
Some more info: Date entered into Regular Classes: October 1, 1995. The Border Collie was recognized by the AKC for inclusion in the Miscellaneous Class in 1955. (AKC web site)
If you scroll down a little more on the same page you will note a breed standard. In short, the breed IS ALREADY STANDARDIZED. Any inbreeding is not the fault of the AKC. It is the fault of clueless and/or idiotic and/or greedy owners generally fueled by the desire to make a quick buck of the popularity of a breed (indirectly aided and abetted by an ignorant public-such as you). Inbreeding is a FAULT. Good (ethical) owners/breeders take great pains to avoid inbreeding as it can permanently damage a breed's genetic diversity and introduce genetic disorders that are extremely difficult to overcome (probably what you think you are referring to...)
"Both lagers and ales are beer. Lager is a generic term for all bottom fermented beers.
A Pilsner, or Pilsener, is a bottom fermented beer, that is, a golden coloured lager with a crisp palate and a dry finish due to the addition of more bittering hops, usually noble hops such as Saaz."
http://www.fmbrewery.com/faqs.htm
Furthermore, the best, and indeed the only way to fix a set of alleles within a breed is through inbreeding....
7. The Standard
The existing Border Collie is not a breed without a standard. It has a very specific standard, by which dogs without registration papers and pedigrees can be Registered on Merit if they can demonstrate their herding ability to satisfy this standard. Whatever appearance standard is designed by the AKC and its chosen Breed Club (should it eventually designate one), it will not be the same standard to which the breed currently strives; it will therefore, by definition and unavoidably not be the same breed of dogs.
Even though the initial registration will come from the existing breed, the next generation of "showdogs" will have been bred under a different set of selective rules, and will already be at least philosophically different. After three years, when the AKC closes its books and no longer allows dogs of the original breed to be used for breeding, the AKC breed will have become a separate entity, no matter what its name!
This already happened at least once, when the "Lassie" collie was created. The working sheepdogs used to be called "collies." They became "Border Collies" to distinguish them from the developing show breed. At the time of separation, there was no real distinction; anyone can tell the two breeds apart now.
All of this is quite apart from the possibility of a standard being chosen which is simply inconsistent with the demands of the shepherding life. This may be in the written standard or in the fashions of judges who know nothing about these physical demands. This has already happened to some of the breeds (Labrador retrievers, for instance, are currently too heavy and short-legged to be of much use in the field; Siberian huskies tend to be showring winners with legs too short to run properly and with fluffy coats that cannot shed snow and ice; bearded collies look nothing like their ancestors, and have coats which obscure their vision, and collect burrs and mud). There has been some call for the USBCC to become the breed club so that we could set the standard and thereby avoid the problems of inappropriate physical traits being used. Unfortunately, although the problem will be made worse by the "wrong" standard, it is the existence of a physical appearance standard, and not its details, that is the danger. The currently proposed standard is flexible enough to appear to cover many of our dogs. In practice, however, an appearance standard, however broad it may seem, will subject the breed to all the problems listed above.
Although there is a popular belief that a dog that looks like his father (or mother) will work like his father (or mother) this is simply not necessarily true. Because of recombination of genes, it is no more likely that the pup with his father's markings is going to behave more like his father than the pup with completely different markings. If we were to set the show standard to duplicate in every detail the appearance of the latest International Supreme Champion, this would no more guarantee us a working breed than any other conformation standard. If we don't choose the pups that work like the latest Champion, we are not selecting the right genetic blend from the many possible combinations.
8. What Is A Breed?
As was stated in the USBCC Spring Newsletter:
"To a geneticist, a breed is simply this: a population of animals whose breeding is controlled and outcrossing limited, so that genetic selection can be exercised on it. . . . A population is simply a subgroup of the whole species of dog, Canis familiaris. Controlled breeding and limited outcrossing make it possible to select . . . for whatever genetic traits the organized breeders decide on. Organized breeders is almost a necessary part of the definition; one breeder cannot produce enough dogs to truly create a breed, and a lot of breed
Seastead this.
pilsener is czech. as is budweiser.
read up.
All pilsners are lagers. But not all lagers are pilsners. Get it straight, man.
Constitutional rights may be respected, repealed, or modified; but they must never be ignored.
And don't forget, those that speak two languages have roughly double the vocabulary of someone that speaks only one
I don't buy that. I live in a very racially diverse area, and have a number of friends and acquaintences who learned English as a second language.
The age at which they learned English varies from early childhood to adolescence, but one thing they have in common is that their vocabulary in either language is not as good as a native speaker's. These are my friends, so don't take this as some sort of insult to people who speak English as a second language -- this is something they freely admit.
In general, their conversational vocabularies are perfect, just as large as a native speaker's. But there are a tremendous number of words, often obscure or technical, that they know in one language but not the other. A Chinese friend of mine, for instance, told me that she has a lot of trouble talking to her Chinese-speaking friends about computers, because she only knows the technical terms in English. And Chinese is her native language. I would guess that I know as many English words as she knows of English and Chinese put together. Judging from what I have seen, I would guess that that is pretty representative of the average bilingual person.
Obviously some bilingual speakers will have an average vocabulary in each language (and therefore double the average single-language speaker's), just as some people who only speak one language have double the average person's vocabulary. But I don't believe that that is the general case -- people can only remember so many words, and branching off into another language doesn't magically make your memory bigger.
ZFS: because love is never having to say fsck
Learning and being able to understand a large vocabulary of words is a only a subset of what language is. The ability to use grammar to string together ideas is a core fundamental of language and only utterances that have this characteristic are real languages (as opposed to pidgins, etc).
If a dog or any other non-human creature could learn to correct sentances like 'I forgotted the name of my neighbor' (or any such novel, grammatically incorrect utterance) then that would be a real discovery.
A number of breeds have reputations for intelligence, and the border collie is certainly one. Others I'm aware of are Doberman pinscher, German shepherd, standard French poodle. An Australian shepherd is said to be almost uncanny in its intelligence.
Contribute to civilization: ari.aynrand.org/donate
We originally embarked on the quest for teaching apes language because we felt that it would be a good indication on our successfulness to speak with aliens if we ever found them. To communicate with sentient non-humans we would need an ambassador that could speak their language.
So we grabbed our closest relatives. The apes. There is plenty of proof they have some intelligence, for example wild chimpanzees will not allow incest within their social circles.
We tried to teach apes how to speak our language. We kidnapped babies and raised them like human children. We forced their mouths around words and were able to teach them a couple of words at best. The chimpanzees simply couldn't learn our spoken language. We were stuck.
Then came ASL. We began to teach apes ASL. We were much more successful with this. We could now communicate with another species. So we had them interact with non-sign language speaking apes. But it was a failure. The sign-language ape knew as much about wild apes as ourselves. To the "speaking" ape, the non-speaking ape was a wild beast whereas itself was an intelligent beast, like humans. When asked to identify themselves speaking apes will identify themself in the human category, rather then the ape category. They identify themselves in a category. To me, that is more then enough proof apes (of the kind that have shown this quality. I don't know which species does this) are self-aware.
I didn't know that they did it this way. I am not as impressed as I was before. The dog is going to realize which one is out of place just by the smell of the toy which obviously doesn't fit w/the rest. Trained dogs sniff out stuff that they recognize all the time. What's so different about them picking the one thing that is different?
You're talking about two different things there. You (a human) make the mental leap between "unrecognized object" and "unknown sound" very easily. That's one thing. Many animals can be taught to recognize various objects or smells. That's another thing. Reasoning that a new (previously unknown) object/smell might be related to a new (previously unknown) sound takes a slightly higher level of intelligence. Or so I would think.
Sniffer dogs are trained on known scents and trained in exactly what to do and how to react to those known scents. They aren't taught to react to something just because it's different. Dogs are so sensitive that there are "different" objects around them literally every time their handler comes to work (new clothes with smells of unknown people or pets on the clothes, etc). This dog does appear to make a sort of minor mental leap.
Your guesses don't make much sense. To prove this, lets say your Chinese friend knows all the common words in both languages, technical/scientific words only in English, and cultural/arts words only in Chinese. For you to know as many words would require a whole new category of words to catch up with all those doubled common words.
A friend of mine who learned Japanese was complemented - even before he was good - on the size of his vocabulary. The reason for this is that, not knowing which symbols to learn, he learned them all - including the rare ones. I don't know enough about Chinese or Japanese to tell you more about that though.
Back to size of vocabulary within a language - there is a major difference between active and passive vocabulary. Add homonyms and proper names and the issue becomes even more confusing. Take the words 'homonym', 'bar', 'Reagan' and 'London'. Homonym is in my English active vocabulary, but I wouldn't know it in any other language I speak (despite being fluent in another and able to communicate in a couple more). I can think a lot of meanings for bar, ranging from chocolate to law - can the dog deal with that? These can be hard for new students of a language to deal with too - especially when they use the wrong one! Reagan - or any name - is a word too. And London becomes Londres in French, so there is no guarantee that you know all proper names you know in all languages you speak.
Learning a word doesn't take a lot of memory. It is learning the meaning which takes the real memory, and then associating it with the word, and keeping the association - which means using the word at least occasionally.
So for sake of argument I may have a very slight reduction in my 'uncommon' english vocabulary, because I don't live in an english-speaking country. As it happens, I have a much larger rare vocabulary (though much of it is passive), but we shall ignore that and say I have 95% of normal vocabulary in English. By your logic, because me memory is limited - I could only know 5% of my second language's average vocabulary. I would estimate it as closer to 80%.
My numbers and odds are picked out of the air in the following:
The first 100 words are essential and anybody claiming to speak the language at all will know them all (100%).
The next 1000 words are common and anybody claiming to be fluent will know almost all of these (99%)
The next 10,000 words are uncommon and fluent speakers will know quite a lot of them (90%). Bilinguals may know a little less than expected here.
The next 100,000 words are rare and fluent native speakers will - on average - know 50% of them. Relatively fluent learners will know less - perhaps as little as 10%.
All other words are very rare or jargon.
So the main area of opportunity for increasing the total size of your vocabulary is in the rare area. But this is where the words you rarely need are (unsurprisingly!) Whereas learning the common and uncommon words in a second language is easy if you are using that language quite a lot too.
Now I grew up in the country of my second language, but never suffered at the hands of their educational establishment. Nevertheless, I am fluent with the exception of some idioms - I know all the words, but not the expression. However, I do admit to not knowing as many rare terms. After all, all my technical reading and study has been in English. However, there are some rare terms which I actually don't know in English - cooking ingredients, for example.
So lets say I know the first 10,000 in each language, but 'only' 45,000 and 15,000 rare words. The 5000 I lose from English I instead know in the other language, and there is an overlap of 10,000 which I know in both. So I know the same number of concepts as the average monolingual person, at the 'cost' of 20,000 vocabulary spaces. But if I hadn't learned that other language, it wouldn't make my English vocabulary 20,000 words richer.
Ah well, that was a lot of stating the obvious.
Ah, a *border collie*... the /. blurb said "collie" which is a different level of dog entirely... about 4 years difference in ability, if they were human children. (Damn article didn't want to come up earlier, either.) Border collies are fairly sharp, if prone to be obsessive; collies compare unfavourably to a box of rocks. :)
:)
Yes, speaking as a professional dog trainer (35 years, and with 11 generations of my own Labs to date) that's exactly right: We have to select for desired behaviour (in your example, herding drive, "eye", and the ability to "balance" the flock) and intelligence (the brains to figure out how to help sort, bunch, move, manage, and protect the flock, instead of regarding them as an easy lunch: herding is fundamentally *interrupted hunting behaviour*).
These abilities are all inherited, thus subject to selection. If they weren't, why bother? Why not breed sheepdogs from parents that think lambs are a tasty snack?
More on the language-understanding thing in my posts further up. But yes, it is easier to get a complex job done with a dog that has a good broad understanding of language. (In sheepdogs and retrievers, that usually means whistles, hand signals, and a couple levels of voice commands.)
Having finally got the damn article dragged across... it's not unusual for a bright dog to learn that this here word == that there object in only one or two exposures.
~REZ~ #43301. Who'd fake being me anyway?