Strange Loops

When you think of intelligent non-human animals, you probably think of apes: they use tools, appear to have culture, can be taught language-like communication systems, and the list of uniquely human traits seems to be ever-shrinking thanks to them. Maybe you include dolphins in your list of smart animals.

When asked to imagine intelligence in the animal kingdom, it’s unlikely that the critters coming to mind would be octopuses, fish, birds or insects. But as a recent Scientific American article on comparative cognition points out, some branches of the tree of life that don’t normally get a lot of attention for their smarts have actually demonstrated some pretty impressive abilities.

In the sea, cephalopods (octopuses, squid, cuttlefish) are the mental badasses of the invertebrates. Octopuses, for example, are not only great problem solvers, but can learn to solve a problem simply from watching other octopuses do a task. Fish may be smarter than we give them credit for. Goldfish, for example, can orient their way through mazes (more efficiently than slime-mold, even). So can reptiles like turtles.
(more…)

Despite our many lapses, humans still manage to show remarkable self-control. We pass up a tempting slice of cake in order to eat a healthier alternative. We avoid buying a shiny new car today so that in a year we can put a down payment on a house. We save for retirement. Sure, we may not be perfect at avoiding temptation in the present, but when you think about it, the amount of self-control we do show is rather impressive.

Some people are better than others, of course. In the 1960s, Walter Mischel tested young children by giving them a marshmellow. They could eat it immediately if they wanted, but if they waited 15 minutes, they got a second marshmellow. Some kids succeeded in waiting, some didn’t; and it turns out the ability to wait was linked to success later in life.

Self-control is a valuable skill, but is it unique to humans? We can look at our close evolutionary ancestors, the non-human primates, for a hint.
(more…)

Short answer: yes, it is irrational. Odds of winning big are less than the ratio of ticket cost to amount won. If $1 has a 1 in 1-billion chance of netting you 500 million dollars, it’s a really bad deal. Why?

Let’s say you could play the lottery over and over and over again an unlimited number of times. After a trillion plays, on average you’d win about 1000 times (roughly once every billion draws). That’s 1000 x $500 million = $500 billion won, but you’ve spent a trillion ($1000 billion) to play. So you’ve wasted a lot of money in the long run. Even if the chances of winning are closer to the cost and win amount ratio, if the odds are lower than the cost and win amount ratio, then it’s generally a bad deal.

So traditional economics says not to play the lottery.

It’s hard to argue with the simple probability of such a straightforward game, but unspoken behind the conclusion to not play lottery are a number of assumptions to which those probabilities are applied.
(more…)

At the Primate Research Institute in Japan, Ai is a chimpanzee in her thirties who has been involved in cognition research for decades. She’s well-known for learning to use our familiar numerals (1, 2, 3…) to appropriately label sets of objects (5 bananas, etc.), and she’s done some other really cool things with numerals and numbers.

What I didn’t know until recently is that she has a son, Ayumu, and he too is becoming quite the research superstar. A recent study demonstrated his talent for remembering an array of randomly mixed numerals on a computer screen, after they were very briefly flashed to him and then masked by white boxes. He had to touch each of the hidden numerals in order without a single mistake in order for a trial to be considered a success. This seems to show a pretty impressive working memory for visuo-spatial layouts, but does that mean Ayumu has something like eidetic memory (”photographic memory”)?

I don’t have the full article. However, previous studies with other chimps (including Biro and Matsuzawa’s 1999 study with his mother Ai) have shown that they “queue up” action sequences in a task like this, such that if the targets change/move in the middle of a trial, their fingers still briefly travel toward where the right answer would have been. Perhaps this is a demonstration of a pre-planned motor sequence, rather than anything like photographic memory.

One way to test this might be to remove one or more of the number locations completely (including the white box masking where it originally was flashed). If the chimp has a photograph-like visual of the layout in memory, then he should be able to continue the sequence by skipping over the missing numeral. But if he has pre-planned his actions in advance (as mother Ai did for a set of 3, 4 or 5 numerals), then his finger should still move to where the missing numeral disappeared.

Either way, if the task sounds easy, try it for yourself. (even with only 5 numerals, it’s pretty tough at first). The study compared college students’ performance to that of Ayumu, and the chimp won. Note that without access to the paper, it’s hard to know if the setup was truly the same. Ayumu almost certainly had a lot more practice than the college students (but then this is necessary to teach him what the task is, whereas humans can have instructions explained to them). Also, Ayumu’s numerals aren’t masked until he touches the first one (usually very quickly), but the college students’ might have been masked immediately after flashing (as the website demo with fewer numerals does).

At any rate, the experiment is a very impressive demonstration of a larger memory span than has been seen before in a setup like this. However that information is stored, it’s cool that he can do it.

[Thanks to Primatology.net for the game link]

The Telegraph (Dec 2007) has a story on Liz Hawkins’ research into dolphin communication which claims they might be using language.

Basically, they used different whistles and clicks depending on environmental and behavioral context. Not exactly a surprise: many species do this. Vervets monkeys are well-known for making distinct alarm calls to their fellows based on the type of predator spotted. But is this enough to call it language?

Linguists have identified a handful of general properties of human language which give it its expressive power. Some of the crucial ones:

• Arbitrariness: the sounds, letters or hand-signals used do not directly resemble the objects or ideas they refer to. The word cheetah does not look or sound like a big cat, and a vervet’s alarm call when eagles and hawks are around does not sound like a bird.
• Productivity: can create new strings (combinations of sounds or words) from smaller pieces, i.e. can say something that has never been said before, and people will understand.
• Displacement: allows reference to the past or future, or things out of immediate sight/experience. We can talk about yesterday’s weather and events that never happened.
• Duality: has two levels — meaningless sound pieces (phonemes) and semantic meaning (morphemes, words, etc.) — operating at the same time. That is, mouth sounds like the ‘b’-, ‘a’- and ‘t’-sounds in “bat” mean nothing alone, but in the proper sequence they refer to an animal.

Whether all of these (and other important properties) are necessary to classify communication as strictly linguistic, it is clear that what we normally mean by language (i.e. human-like communication) is much more complex than simply producing different sounds or signals in distict contexts. So they’ve got a long way to go in establishing that dolphin communication is a “language” in anything close to the same sense by which we apply that term to human communication.

Certainly, other researchers have done work investigating dolphins’ communicatory abilities (including linguistic precursors like equivalence classes), but as of yet, most scientists in the field would probably not be comfortable calling it language. That’s one problem with popular media science writing: the writer tends to go for catchy, succinct headlines and descriptions that at best over-simplify scientific results, and at worst present outright falsities. It is important to spread scientific knowledge, but we must also spread scientific literacy in order for that knowledge not to be corrupted.

At any rate, the Telegraph article ends with a quote from Liz Hawkins: “Dolphin communication is much more complicated than we thought.” Fair enough, but perhaps she should replace the word ‘complicated’ with ‘flexible’. Making different whistles when feeding versus traveling is not exactly complicated.

There’s still a lot of research to be done, but this is a really exciting area because dolphins provide a rich source of data from an evolutionary line rather distinct from primates (the target of most animal language research in the last half of the twentieth century). Nice to see that field work is complementing the controlled experimental work.