June 26, 2009
Dreaming of Nonsense: The Evolutionary Enigma of Dream Content
From Scientific American:
Friday, June 19, 2:12 a.m.: Loading up the trunk of my car with clothes hangers when approached by two transients… try to engage them in good-natured conversation about the benefits of wooden clothes hangers over metal ones, but they make me uneasy, say they want to go out to get a drink but I’ve got to go. In a city somewhere… looks like a post-apocalyptic Saint Louis.
Saturday, June 20, 4:47 a.m.: Was just now trying to return my dead grandmother’s cane to her. Took elevator to her apartment… meant to go to the 8th floor, but elevator lurched up to the 18th floor, swung around violently then shot back down. Could hear voices in the corridors outside elevator shaft…. a mother yelling at her child. Grandma then became my other grandma, also decesased, yet in a nursing home; doctors say she’s doing fine.
Sunday, June 21, 5:02 a.m.: On a floating barge in the sea trying to get to some other country, just made it, the dogs are running all over the place but seem more like rodents.
Monday, June 22, 3.31 a.m.: Just learned that one of my colleagues died suddenly, everyone’s in shock (they say it was “an accidental overdose of oxygen from a breathing tank; he fell asleep”). Can’t believe it, was just talking to him today about death. Also something about an airplane delay… need to get home but can’t find my test results to submit, searching all over, trash cans, pulling out drawers… people preoccupied.
These are dreams, of course. Mine from the past few days, to be precise—and they are totally absurd. Why on earth do our minds conjure up such ridiculous imagery, such inane thoughts, such spectacularly vivid and surreal landscapes, intense emotions—such narrative trash?
More here.
Posted by Azra Raza at 06:06 AM | Permalink
Comments
"Ridiculous imagery, inane thoughts, narrative trash?" Of course these examples of Bering's recent dream life seem somewhat orphaned when presented out of context with details of his waking life, which he doesn't provide.
Dreams are metaphoric landscapes, offering a haze of either/or, multiple doorways to personal revelation. Unfortunately, they will always seem inane to Bering until he learns to read himself in this particular way.
Some scientists, like religious fundamentalists, have a horror of metaphor. Perhaps he should read the work of Bert O. States, particularly The Rhetoric of Dreams, and Dreams and Storytelling. Wouldn't hurt to examine the neuroscientist Rudulfo Llinas' work on the intralaminar nucleus, as well.
Posted by: Philip Graham | Jun 26, 2009 2:23:51 PM
"Dreams are metaphoric landscapes"
This doesn't really help much in explaining why we dream, which is what the article was addressing. Or maybe you think that is the purpose of dreams, to send us metaphorical messages?
We do have some understanding of why dreams have the form they do, but we don't know much about the content. The problem with the metaphor theory is that it's unfalsifiable, and therefore basically meaningless. If I gave you one real dream I had and one dream I made up, could you tell which was which? Or, what if you misremember your dream, would the metaphors still hold up?
The metaphor theory tells us something about how we get meaning from things, but nothing about the things--dreams--themselves. You'd probably get the same result by getting out a photo album and flipping to a random page.
Posted by: billy | Jun 26, 2009 5:39:58 PM
Yes, dreams are indeed unfalsifiable, which is why "scientific" approaches are often (but not always--see Llinas' work) so ham-handed.
Why would I care whether you made up a dream or not, and why would you bother to do so? Your dreams are your own. If poor Bering thinks his are inane, then he will get nothing from them. Why live the "unexamined life"?
By the way, dreams don't send "us" metaphorical messages, because they are not outside of us. Everything in a dream is self-generated, they are our own individual messages, subject to our own psychological context, our own individual narrative grammar.
Posted by: Philip Graham | Jun 26, 2009 6:42:34 PM
I didn't say dreams were unfalsifiable, I said your metaphor theory was. We already have some sense of what the forms of dreams are. For example, smells are rare, so if my "dream" was peppered with lots of them, there's a good chance I made it up. If we do settle on some theory about dream content, then we will be able to say whether a given dream is genuine or not (or give a probability).
The question here is whether there's anything special about dream content--is it random, an artifact of some other process, or does it serve a purpose? Unfortunately the metaphor theory doesn't distinguish between these possibilities so ultimately it's scientifically uninteresting.
Posted by: billy | Jun 26, 2009 8:11:28 PM
It's uninteresting to try to apply scientific reductionism to dreams. Unless it can take into account deeper individual context, and the interpenetration of waking and dream life, and a full understanding of the fluidity of individually generated metaphors, this method is as doomed with dreams as it is with trying to create true AI.
As Llinas' research on brain structure implies, the human mind is a "dream machine," relatively constrained or let loose during the 24 hour daily cycle. And as John L. Caughey's research on daydreaming, Imaginary Social Worlds, shows, we are indeed "away," in a modified dream state, for a significant majority of our waking lives.
When I read certain folks' assertions that their dreams are nonsense, I feel for them--they appear to have such limited awareness of or access to their inner lives.
Posted by: Philip Graham | Jun 26, 2009 8:53:35 PM
Maybe you could give some links?
Posted by: billy | Jun 26, 2009 9:18:00 PM
Billy:
Isn't the metaphor theory addressed in the article under "Dreaming as Problem Solving?
By the way, I found the "Dreaming as Costly Signalling" theory the most amusing EP Just So story I have ever heard:
So, want to get hot chicks? Fall asleep in public!
Posted by: Vicki Baker | Jun 26, 2009 10:06:50 PM
Very funny, Vicki. If we have but one purpose, evolutionarily speaking, then everything we do during the reproductive window is costly signaling. It certainly feels that way now. I am reminded of the Kurt Vonnegut novel -- _The Sirens of Titan_? -- in which it is discovered that, seen from a suitable distance in Space, the Great Wall of China forms one gigantic ideograph whose meaning is "Hello!"
Posted by: Elatia Harris | Jun 26, 2009 11:03:17 PM
Freud's Ego is unfalsifiable too--not to mention scientifically unobservable. And yet here we are...
Hardliners are so no fun.
Posted by: Lambness | Jun 27, 2009 2:31:21 AM
"Also, many of the awkward, embarrassing, anxiety-producing experiences from our more negative dreams tend to filter into our waking life, leaving a sort of lingering emotional residue that puts us at an adaptive disadvantage by compromising our everyday social interactions."
You mean "we" find ourselves continually checking to see if we are still wearing pants?
Posted by: Carlos | Jun 27, 2009 7:39:25 AM
Billy:
Here's the best I can find on the Caughey book, which is obscure but quite brilliant:
http://books.google.com/books?id=Vi23AAAAIAAJ&dq=John+L+Caughey&source=an&hl=en&ei=Tj5GSt65JYqONqCQrZcB&sa=X&oi=book_result&ct=result&resnum=4&pgis=1
As for Rudolfo Llinas, here's a NYTimes article on his work, "Listening to the Conversation of Neurons":
http://www.nytimes.com/1997/05/27/science/listening-to-the-conversation-of-neurons.html?scp=8&sq=Rudolfo%20Llinas&st=cse
Finally, Bert O. States. Here are a couple links worth looking at:
http://www.anth.ucsb.edu/projects/esm/IAM/BStates.html
http://www.asdreams.org/journal/articles/13-1_states.htm
Posted by: Philip Graham | Jun 27, 2009 11:56:18 AM
Oh, and here's another excellent link for Bert O. States--Seeing in the Dark, Reflections on Dreams and Dreaming:
http://cogweb.ucla.edu/Abstracts/States_97.html
Posted by: Philip Graham | Jun 27, 2009 12:01:01 PM
Thanks for the links. I think I will read some of their books. However, I don't see how any of this backs up your theories. But maybe I misunderstood what you meant from the get-go. From your last link, States:
"Memory, in short, is a dynamic process of up-dating what we already have in mind, not simply a storage-house for what has already happened. It is only a hunch, but it seems to me that dreams may be our clearest window into this whole process of on-going conversion of experience into patterns that help to maintain order in the system; that the dream does its work without the least awareness that we, as dreamers, are looking in on it all and that the dream is finally beneath our understanding in any more than the self-evident sense in which dreams do reflect our concerns, fears and needs. Beyond this it is anyone's guess as to what is going on in dreams or why such a mechanism is biologically necessary; but for various reasons I will claim that the purpose of dreams has little to do with keeping us self-informed, as if dream-work were an interior alter ego, whispering messages and giving progress reports to the dreamer from some mysterious clearing-house."
It doesn't sound like he thinks dreams are messages of any kind, much less a "doorway to personal revelation". I should have mentioned this simple point earlier, but if dreams are messages for the dreamer, why are they so hard to recall? More States:
"Another of my assumptions is that if dreams mean anything the meaning is no different in kind from the meaning that may be given to waking experience. If you want to find out about yourself or what is "on" your mind, you can do just as well examining what has passed during the day. Curiously, one rarely hears of anyone examining the day -- or even a day dream -- for its meaning. For the most part, we take each day for granted. Who has ever thought to ask, "What did my vacation in Maine mean?" But a dream about the vacation is another matter. We like to believe that dreams are "up to something," that the events they depict are deliberately charged with a significance that the real experience (the vacation) did not have."
This is all suspiciously close to my claim that you'd "get the same result by getting out a photo album and flipping to a random page". It doesn't seem like he thinks that the content of dreams is especially important, from his first link:
"The paper is chiefly concerned with dreams and the possibility that dream coherence--i.e., meaning, making sense--as we recognize it in the waking state may be largely irrelevant to the organization and function of dreams which obviously have no artistic mission to be shared with others (readers, auditors) as forms of communcation... The content of dreams--coherent or otherwise from the waking standpoint--may be less important than the act of dreaming itself."
I don't think I disagree with anything he says, so I must have misunderstood what you meant. Maybe you could explain more?
Vicki: If the problems are solved by way of metaphor, then perhaps. But it might be that the brain is reorganizing itself (and solving problems), and that the images or stories we see in our dreams are only incident to this, only the surface of the inner workings of the brain. Then we might have answers to our problems when we wake up regardless of whether the answers rose to the level of "dream consciousness". Of course there's lots of evidence that sleep helps us solve problems--but it's not clear whether dreaming is vitally involved in the process. Usually the types of problems that are claimed to be solved by revelation in a dream are big ones--ones that are on our minds during most of the day. So it's only natural that our dreaming minds would obsess over them as well, so it could just be coincidence that once we found a solution that it'd show itself in a dream (concretely or metaphorically).
In any event, I think the metaphor theory is much broader than than "problem solving"--indeed, isn't one of its selling points that it alerts us to problems/issues/feelings that we weren't aware of beforehand? But I get the feeling that what I mean by metaphor theory is different than what others mean by it... (I agree with your opinion on "costly signaling theory".)
Posted by: billy | Jun 27, 2009 5:40:33 PM
So, want to get hot chicks? Fall asleep in public!
To get the elite babes you have to do more than just sleep; you also have to twitch your eyes and mutter nonsense. Sex magic!
Posted by: Chris Schoen | Jun 27, 2009 5:51:04 PM
Suzanne Langer, writing in 1946, on how dreaming doesn't jibe with the computational theory of mind (though I don't think they called it that back then):
And later, on the metaphorical quality of dream imagery:
Psychoanalytic explanations of dreaming aren't "irreconcilable" with Darwinian theory; they just focus on the aspect of meaning rather than fitness. Even if the origin of dreaming is in something like "brain conditioning" (which is plausible, given that species like dogs, who don't conceptualize, also appear to dream), it doesn't follow that the content of dreams have no semantic meaning. This is an example of the "genetic fallacy," where all explanatory power adheres to a thing's origin.
Incidentally, Barrett's preferred "problem solving" theory is almost identical to the Freudian one she discards as unscientific. The only real difference is that Freud suggested this problem solving needed to be shielded from the conscious mind, because its content was too emotionally powerful to integrate (which was his explanation for why dreams seemed to be narrative garbage.)
Posted by: Chris Schoen | Jun 27, 2009 6:50:03 PM
Interesting:
http://web.mit.edu/newsoffice/2001/dreaming.html
Are they making a leap in concluding that because the same neurons are firing in the same pattern that the rat is having the same thoughts?
Posted by: Carlos | Jun 27, 2009 9:40:36 PM
Sounds that way, Carlos:
"We looked at the firing patterns of a collection of individual cells to determine the content of rats' dreams. We know that they are in fact dreaming and their dreams are connected to actual experiences."
I'm assuming they were able to do some kind of baseline correlation of neural activity to dream content by actually interviewing the rats themselves, right? I mean, that's just good science.
Posted by: Chris Schoen | Jun 27, 2009 10:53:55 PM
Worse than that Chris, they couldn't get a baseline for the rats' waking thoughts! I mean, just because a rat's navigating through a maze doesn't mean that she's thinking about it. These scientists should have taken some time to talk to the philosophers before starting their experiments, then they would have realized how hopeless it is to try to figure out whether animals have thoughts, much less what those thoughts are about.
Anyway, pretty amazing stuff, and that was eight years ago.
Posted by: billy | Jun 28, 2009 2:18:28 AM
Ah yes, it's a widely-known fact that the complexity of a rat brain and a human brain is equal--so, problem solved. I'm guessing that's why the philosophers you refer to would be . . . rat philosophers? I'm not familiar with their distinguished body of work, perhaps some references could be provided.
Being and Cheese, perhaps?
Posted by: Philip Graham | Jun 28, 2009 9:09:26 AM
We have to make some assumptions about how rats work, like that they have thoughts and that those thoughts are reflected in their brain. Of course we can't prove this, but we can't even prove that other people have thoughts--that's just the nature of subjective experience. So it seems to me that the only leap the scientists are making is this one, and without it there wouldn't be a question worth investigating.
And it's not that being able to talk to rats would remove all doubts about what they're thinking. People can and do lie (purposely or not)--we don't have an objective view of our own thoughts. For example, with split-brain patients, the right hemisphere will do an activity and the scientist will ask the left hemisphere (the one that can talk) why she did that, and she'll come up with some completely bogus explanation, because the two hemispheres can't talk to each other. So there's no guarantee that what a person says actually reflects what they're thinking. (This is especially bad for dreams, which aren't dependent on rules of physics. For example, if I recall a memory about leaving the house and then being at the store, I know that I drove there and it took about ten minutes, even if I don't remember those specific details. But if this happened in a dream, then I can't make that inference--our perceptions of time (even the order of events) are unreliable.)
But we do have lots of evidence that our thoughts are reflected by patterns in the brain--we do the same actions and we see the same patterns, and we can predictively map certain types of thoughts to specific regions of the brain. In sleep there's very little outside movement to provide independent clues about thought, but it's not unreasonable to conclude that if the same patterns are occuring in the brain, then the same thoughts are being thunk.
By the way, Philip, did you have a comment on my last post, before the discussion went off on this tangent?
Posted by: billy | Jun 28, 2009 3:48:13 PM
Sorry--didn't have a comment on that--I'm traveling now, not always online, and pretty busy.
But your line "we can't even prove that other people have thoughts" is both amusing and alarming to me. Books prove that other people think, as does music, art, and so on.
We are indeed hard-wired for isolation and solitude--we can't hear other people's thoughts, and speech, as you note, is notoriously unreliable. Art is one of the great, hopeful attempts by the human species to bridge that gap.
In fiction, for instance, the simple phrases "I thought," "she imagined," or "he remembered" offer readers access to an inner world not their own. Such inner revelation is based on a writer's understanding of his or her own thoughts and imaginings. Depending on the artistry of the writer, a reader is able to accept those depictions, because they reflect to a certain extent his or her own thought patterns.
Dreams are harder to get across in fiction, because of the inherently slippery nature of their imagery and narratives, but there are masterly examples throughout literature, from Tolstoy to Updike, etc. But just because something is difficult doesn't mean writers shouldn't try.
Finally, I don't think that we ever "do the same actions and we see the same patterns." I'm with Heraclitus--"You could not step twice into the same river; for other waters are ever flowing on to you." There are always variations and additional subtleties at work in every moment. José Saramago has a lovely phrase for these subtleties: "subgestures." If you're really serious about examining the world, he says (I'm badly paraphrasing here), don't settle for the banner headlines, go for the fine print.
Posted by: Philip Graham | Jun 29, 2009 12:12:37 AM
I agree that other people have thoughts, although I would first point to our similarities in appearance and behavior, and then cite language, that we can directly communicate with each other and share our thoughts, before I went to more abstract modes of communication, like books, music, and art. And surely rats don't have all these things, but it seems rather evident from their actions that they have thoughts. (I'm under the assumption that you believe that rats do have thoughts, even though our brains are more complex than theirs. If that's wrong, you should let me know!) I only introduced that philosophical argument because I thought that that was the only place where there were gaps in logic. Once we accept that thoughts exist, then we can review the science and conclude that the brain is where those thoughts are generated. When we see the same patterns in the brain, we can conclude that the thoughts are being thunk. What's the alternative? If the same patterns occur but different thoughts, what accounts for the difference? What are we missing?
I couldn't tell whether Carlos and Chris were just trying to be funny, or if they also had a real problem. So I tried to first take the argument to its extreme, and then tried to explain why the assumptions the scientists made were reasonable ones.
Finally, even if the patterns weren't exactly the same, they were evidently close enough so that the scientists were able to identify them. I didn't look closer at the research, so I don't know for sure.
Posted by: billy | Jun 29, 2009 2:59:44 PM
Speaking for myself, I wasn't kidding. Brain activity can indicate different things, but we can't tell by looking at a brain scan what the precise thoughts are, only general activity (such as problem solving, language parsing, anticipation, etc). Whether or not the rat was precisely replaying the actual event, running (in a meadow, in a castle), or even (being trained to run mazes for a reward), simply anticipating a snack.
One way to tell would be to have had the rat run 3 different mazes, and see if they could tell which maze the rat was running in his dream.
Posted by: Carlos | Jun 29, 2009 3:26:21 PM
I'm under the assumption that you believe that rats do have thoughts, even though our brains are more complex than theirs. If that's wrong, you should let me know!
I think you are wrong. I think that rats have feelings, and inclinations, and reactions, and urges, but to have "thoughts" requires some kind of medium of reflection, like a language.
Maybe you could give an example of what you think a rat "thought" might be like, from the rat's POV?
Posted by: Chris Schoen | Jun 29, 2009 4:22:10 PM
Like a movie? Would you need language to recall the visual/olfactory/kinesthetic details of running & finding cheese? My dreams don't seem primarily verbal; there are conversations, but it's very visual, mostly.
I watch my dogs "running" while they sleep, and imagine that they are rehearsing/replaying hunting behavior. I suppose though that the brain could just be optimizing a sub-routine, or even just running a screen saver.
Posted by: Carlos | Jun 29, 2009 4:50:56 PM
Chris, what about when a rat navigates through a maze? Can you reduce that process to just reactions and urges? Are we thinking when we run through a cornfield maize? Is there anything fundamentally different between our thought processes and the rats' in these situations? It seems to me that whatever you want to call a "thought", that they're occuring when we run through a maze.
Carlos, the scientists were able to determine their precise thoughts! From the article:
"The correlation was so close that the researchers found that as the animal dreamed, they could reconstruct where it would be in the maze if it were awake and whether the animal was dreaming of running or standing still. These memories were replayed at about the same speed that the animal had experienced them while awake."
Although I do like your three maze experiment. Some links I found, that seem relevant:
Scientists selectively erase fear memories and gain insight into how the memory works: http://www.physorg.com/news156012310.html
Decoding short-term memory with fMRI: http://www.physorg.com/news154466439.html
Mind-Reading Tech May Not Be Far Off: http://www.popsci.com/scitech/article/2009-06/mind-reading-tech-way?page=
Posted by: billy | Jun 29, 2009 6:39:35 PM
Billy,
I am not saying there is no crossover between mice and ourselves or any other organisms with whom we share a lineage.
But there are aspects to the mental experiences humans have running through a maze that rats cannot experience. I personally denote those aspects with the word "thoughts," by which I mean concepts.
We are talking about the difference between symbolic and pre-symbolic mentation. It's a big one.
Posted by: Chris Schoen | Jun 30, 2009 1:23:13 AM
I don't know, Chris. Mammal brains are all structured similarly, it's just that some regions of our brains (e.g. cortex) are much larger. Does this mean we think in fundamentally different ways, or is it just a matter of degree? I tend to lean toward the latter, but if you define "thoughts" to be the things that only we can do, then there's not much I can say.
It's hard for me to separate language and "concepts", but even if I didn't have a word for "mom", I think I'd still have a concept/symbol of her. And I suppose the same would be true for rats, don't they have social lives? Not to mention all the objects in their lives and the way they manipulate and navigate their world. What exactly are the concepts that humans use when running through a maze that rats don't use? If you plugged some wires into my brain and let me run through a maze a few times, do you think the patterns in my brain would be fundamentally different than those in the rat's?
So, does this mean that you think the research in Carlos's article was pointless? Or are you fine with it as long as they don't call whatever the rats are doing while awake or asleep "thinking"? I know rats aren't pondering abstract things like "freedom" (or are they?), but then I wouldn't say that that's all I mean by "thoughts".
Here's an attempt at getting at the rat's POV: http://www.ratbehavior.org/perception.htm
Rats Laugh When You Tickle Them: http://www.youtube.com/watch?v=myuceywaOUs
Metacognition: Faced with a test, rats can check their knowledge first: http://www.eurekalert.org/pub_releases/2007-03/cp-mfw030607.php
Have a comment on this last article?
Posted by: billy | Jun 30, 2009 4:59:00 AM
but even if I didn't have a word for "mom", I think I'd still have a concept/symbol of her. And I suppose the same would be true for rats, don't they have social lives? Not to mention all the objects in their lives and the way they manipulate and navigate their world.
The only way I can respond to this is that there is evidence for concepts in human thought, and that is language. There is evidence of very crude symbol use among great apes, with human intervention. There is something going on that could be language among cetaceans. That's it.
As far as we know, other organisms use signs, not symbols. At least there is no evidence of other organisms using symbols. Without them, I am at a loss to see what the currency of concepts could possibly be.
Amoebae and bacteria "manipulate and navigate their worlds" as well. Should we assign them the faculty of conceptualizing too?
In short, yes, I define thoughts as something only conscious symbol users can have.
Posted by: Chris Schoen | Jun 30, 2009 8:03:23 AM
As someone who actually studies animal cognition:
(That opening remark to be read either in as unpretentious or as-pretentious-as-possible a tone, to your taste).
'Thought' isn't going to get you very far as a concept--which is why no one studying in this field uses it. Is that a mental representation, something cognitive like the ability to remember aspects of an object when it is out of sight ('object permanence'--soundly demonstrated in many species of animals)? Or an association, like knowing that a lever will supply food? A declarative rule used to update already-formed memories of past events? Inferential reasoning (e..g., 'transitive inference')? Cause there's lots of evidence of all of that in an array of species, including many birds as well as apes (and even fish!).
As pop-science writers go, Damasio nicely articulates some ideas on basal or core consciousness (as opposed to extended or autobiographical consciousness, which extends core consciousness by adding a temporal component). That idea neatly evades the human-centric (and I think rather unimaginitive) idea that language is needed for experience or 'thought'. After all, if a word merely functions to recall a 'thought', then you don't need words for thoughts! Not many people are impressed by the linguistic story line any more...
A simple and great paper on semantic meaning in language is by Cheney, Seyfarth and Marler, called 'Vervet Monkey Alarm Calls: Semantic Communication in a Free-Ranging Primate', published in Animal Behaviour in 1980. It's a textbook classic so probably easy to find summarized elsewhere. Such a simple little test with such intriguingly suggestive results.
Danny Povenelli writes some good stuff on the tractability of the whole mental representations vs. simple(r) associative learning theory. He's a rather harsh critic of most animal cognition studies, but there's a glimmer of optimism in his writing (which is very lucid).
Also an interesting idea on a possible function of consciousness is to integrate simple associations with cognitive awareness ('Why a Rat is Not a Beast Machine'--read 'automaton'--by Anthony Dickinson articulates this idea).
So this comment has nothing much to do with dreams, no. It is intended less as a warning against drawing premature (and over-confident) conclusions, and more as a hearty recommendation of all the good stuff out there to take this discussion to a higher level (if you're into that).
The best one-stop-for-all recommendation would be a brilliant and comprehensive text(ish) book called, 'Cognition, Evolution and Behavior', by Sara Shettleworth--probably as good a starting point as you could find. It's extremely authoritative and clear.
And I think the commenter above is a *wee* bit hasty in concluding that scientists in this area don't talk to philosophers... just a tad.
Posted by: Kim | Jun 30, 2009 11:23:20 AM
Hi Kim
I had a probably very foolish thought a while ago. I was on my deck at sunset when suddenly one of my biggest trees was mobbed by thousands of small dark birds. They all began simultaneously singing up a storm resulting in quite a remarkable cacophony. My idea was: what if each of them was telling the others how they had done with the day's foraging, and even details on locations, type of bugs found, etc. Like the honeybee's flower dance, but all at once. Is that even remotely possible?
Sorry for the OT, but I figured since I have already successfully derailed yet another thread…
Posted by: Carlos | Jun 30, 2009 1:11:11 PM
With respect, Kim, I don't think any of this overcomes the obstacles I've proposed. Where is the evidence that any of these organisms use "reason" or experience "meaning"? Which of this evidence would not also be explained by simpler behavioristic mechanisms?
It is one thing for a third (human) party to associate neural patterns (for example) with predictable behavior. It is entirely another thing to posit subjective, conscious experience of this type of "reasoning." After all, if modern gene-centric evolutionary biology is correct, this would suggest that genes themselves are capable of reason, which is clearly a fallacy.
I'm not trying to paint a Cartesian picture of animals-as-automatons here. Life is splendidly complex, and in mammals we can recognize a rich emotional life similar to our own. But I think Damasio is absolutely wrong to imagine that non-human animals have a "sense of self, ceaselessly recreated for each and every object with which the brain interacts." Such a sense would require a medium of representation for which there is no evidence. To "detect changes in the self and images of external objects" an organism must be able to compare its perception to a model or representation of its perception. In human culture we can point to intellectual and phsyical models (as distinct from the neural patterns we may display when encountering these models.) What can we point to in the non-human world that serves this function?
I think there is a temptation to view animal cognition as though it worked this way because that is how we understand our own cognition. But without any evidence that non-human organisms manipulate symbols (as opposed to signs, such as the bee dances in Carlos' example), I don't see the grounds for talking about things like "inferential reasoning" in anything but the most figurative of terms.
Posted by: Chris Schoen | Jun 30, 2009 2:29:54 PM
And do amoebae and bacteria have social lives? To me, they seem like a bundle of reactions and urges (although I'm no expert). As far as I'm aware, they don't create an inner model of the world, like rats and humans do. That's what I mean by "manipulating and navigating their world".
It's kind of frustrating to have this conversation with you Chris, since you rarely answer the questions I ask, like:
(1) What exactly are the concepts that humans use when running through a maze that rats don't use? (Since you claim that "there are aspects to the mental experiences humans have running through a maze that rats cannot experience".)
(2) What do you think about the last article I linked to, that has evidence that rats think about their own knowledge?
And another, (3) do you think it's possible for a being to create an inner model of its world without using symbols of some kind?
But now I think I understand you. What you really mean by "thoughts"/"concepts" is abstract symbols that are communicated to others (e.g., language). So the inner representations don't matter--not that you think we could ever definitively conclude this (in rats), since you think that the only way of getting at them is through language. Wait, I'll rephrase that as a question: (4) Do you think there's anyway of deducing that another being has "thoughts" as you define them, except through language? If so, describe an experiment that would do this. For example, if you thought that organism A (say, a human) used "thoughts"/"concepts"/"symbols" when doing activity X (say, running through a maze), and you wanted to see if organism B (say, a rat) also used thoughts, you could let organism B do the same activity X and see whether their were any significant differences in brain activity (since that's presumably where thoughts would be reflected) or performance. If there were not any significant differences, then you could reasonably conclude that organism B also used "thoughts"/"concepts"/"symbols".
And (5), do babies and people who never developed language have thoughts?
Kim, thanks for all the references.
Posted by: billy | Jun 30, 2009 4:19:06 PM
Couple points, in no particular order:
Carlos, there are actually theories that large, mysterious gatherings of birds, both single and mated, which occur around the world, function as informational sharing sessions. That's not to say they're speaking to each other, of course, but that in some way something is learned by each individual. Afraid I know very little about that.
Chris, the animal cognition field is all about trying to find a behaviour that can't be explained with anything simpler. That's INCREDIBLY difficult, as you appreciate. Povinelli would argue it may even be close to impossible. But some instances may exist. For example, being able to understand that another animal believes something incorrectly is regarded as a pretty good indicator that the animal can mentally represent the thoughts of another animal. That's big.
Transitive inference is reasoning by any standard, I would say. Basically, it means inferring from the knowledge of A>B and B>C that A>C. And I don't think we reason consciously, though we can create a conscious report of that reasoning process post hoc.
As to whether or not animals experience a phenomenological thoughts... who knows. That's way beyond our knowledge right now. There may even be some people walking around without conscious experiences so far as I know.
However, the Dickinson article I mention does at least provide logic as to why a rat would benefit from consciousness (he bases it all on an anecdotal experience he had with watermelon once).
It's been a while since I read Damasio, but I don't think the core consciousness he spoke of had anything to do with a sense of self. Rather, it provided the raw feelings that could provide the material that, provided you have autobiographical memory, could be logged away and contextualized into making a sense of self. Taking a neurologist's case-based approach, he noted that people with a very basal part of the brain would report experiences but lack all higher autobiographical abilities if missing other more recently developed parts of the brain. Animals pretty much all share that basal part with us--so maybe they have the machinery for creating those in-the-moment conscious feelings like those brain-damaged people? Granted, this is not strong evidence, but it's about the best we can base anything on right now. I for one think it most likely my dog has consciousness, even if he were to be incapable to think about the past or about me when I'm not in the room.
Do you know the Vervet study? It's suggestive evidence of semantic meaning, as opposed to merely reacting to sounds 'innately' (to use the dreaded word). Basically, vervet monkeys have three different calls for three different types of predators (leopards, raptors and snakes), each of which elicits a different escape strategy. If you play a recording of one of those calls several times, a vervet stops responding to it (crying wolf). Now, starlings nearby also have a call for when they spot a raptor, and the vervets respond to it. Thing is, if you play a recording of the starling raptor call so many times that the vervets stop paying attention, and THEN play the vervet call for a raptor, they still do nothing. They will, however, respond if you played either of the other two calls. This implies that, rather than habituating to the sound, they habituated to... what else? ...the meaning. There are some potential flaws in the study, but I think it's pretty exciting if not conclusive.
We may also be talking past each other, Chris, if your point is that they have an emotional life but no sense of self. I'm not arguing that they have a sense of self. And I think talking about emotion only really matters if it is experienced as a conscious sensation, as we do. Above I gave some reason to think that animals create mental models of things (object permanence is another). That does not imply conscious experience--nothing does at present--but it does suggest 'thinking' about things, and it does suggest reason (there are many studies on insight learning as well).
And I think you might also want to consider how much of our reasoning is actually conscious? When I do actual thinking, it seems more like the answer 'comes to me', bubbling up from below, rather than me 'reasoning it explicitly'.
Posted by: Kim | Jun 30, 2009 4:51:53 PM
they don't create an inner model of the world, like rats and humans do.
I'm still waiting for some evidence of this (for rats.) Where is this model? What is its medium of representation?
(1) What exactly are the concepts that humans use when running through a maze that rats don't use? (Since you claim that "there are aspects to the mental experiences humans have running through a maze that rats cannot experience".)
This is such an open-ended question I don't know how to answer it. Humans can conceivably run though a maze without thoughts (if, for example. being chased, they are driven into a state of such fearfulness that the inner monologue ceases altogether, as is frequently reported after moments of great stress.) Which would perhaps make their experience much like rats.
Or they could be thinking about other mazes they have gone through, and how they hope to perform better this time, or how nice the weather was on the earlier occasion, or what's for dinner tonight.
Or they could recall a maze-running tactic, like "always take the left turn." They could try to visualize the maze from above.
We know all these things (and more) are possible because we have experienced them, or something like them. On the other hand we have no evidence that rats have anything resembling these kinds of thoughts, or could have them.
So the inner representations don't matter.
I don't even know what this statement is supposed to mean. How can you have a representation without a symbolic structure (language) to do the representing? Please don't appeal to brains or neurons in your answer: we are talking of minds here, not brains. Even if one ultimately reduces to the other in the end, we can't skip the step of demonstrating what it is that is supposedly being reduced. Otherwise we are just chasing ghosts.
Regarding your thought experiment, it presumes the very thing that I am denying you have any evidence for: that neural activity can be equated to a specific experience. Even if A and B were two humans, there could be so many contingent factors on what sort of thoughts each might have (personality, personal history, mood, diet, recent events) I don't know how you could ever control for all of them sufficiently to reliably correlate neural activity to mentation.
To extrapolate out to a completely different species, and not a particularly closely related one, seems beyond credulous to me.
And (5), do babies and people who never developed language have thoughts?
The evidence here is close to non-existent, but the testimony of people like Helen Keller and Laura Bridgeman who developed language late, suggests that they do not.
Posted by: Chris Schoen | Jun 30, 2009 5:20:35 PM
Kim,
I recognize that the topic of animal cognition is more complex than I'm giving credit for. But I am troubled by what seems to me to be often unthoughtful, or at least un-thought-out, attribution of human faculties to prelinguistic species. It's not that I'm trying to hold some firewall against the barbarity of apes or anything like that. But I do think that human use of symbolic forms is the primary dividing line between ourselves and other species, much in the same way that animals are different from plants, generally, or vertebrates from invertebrates, and we need to treat it like the radical innovation that it is.
Let me grant that a huge portion of human cognition--probably the majority--is unconscious, and does not avail itself of reason as we commonly understand it. In that sense we have a huge amount in common with rats, vervets, and the whole lot. But this doesn't mean that when we do use symbolic forms like language and logic, that it's not unique.
I don't follow the field closely, but I have never seen any strong evidence for the kind of reasoning you invoke in the form of transitive inference. There's a black box there into which we might be tempted to insert a reasoning process (at least for animals with faces), but this temptation falls short of any compelling scientific reason to believe such a thing.
In the "metacognition" of rats study, for example, the only valid conclusion I see is that rats, quite understandably, have difficulty distinguishing short and long pulses of sound once they are pushed toward an intermediate point. The business about whether the rats "decide" to take the test or shun it based on how well they think they will preform is pure projection, and the rats' behavior is just as easily explained without reference to calculation.
The null hypothesis is that the rats have no metacognition, no knowledge of their own cognitive limitations, and it seems to me this hypothesis predicts the same result. Under this view the rats who couldn't tell the long pulse from the short pulse "shunned" the test to the extent that no strong Pavlovian association was made between the more intermediate bursts and the different levers. Reason need never enter into it.
Studies on humans suggest that we make these kinds of decisions all the time without ever knowing it (which would necessarily preclude knowing what we know!) Signal learning can be accomplished without the knowledge of the subject, as when a researcher positively reinforces (with praise or a smile) responses with an arbitrary component, like plural versus singular nouns in a list, without the subject ever being aware a preference is being established.
"Sense of self" is a direct quote from Damasio's defintion of core consciousness. We aren't even sure that human babies have a concept of self, so it seems a long leap to suggest that other organisms have it just because they share a brain structure.
Regarding vervets, the cross-habituation between starling and vervet calls is interesting, but we're still in the realm of signals here, not symbols. It's perhaps a more intelligent form of Pavloviaan conditioning, and perhaps one that begins to form a bridge to actual symbolic representation, but we need to be careful not to ascribe more to the word "meaning" than is justified. In the world of vervet signs, the alarm call for leopards "means" something like "immediate threat from leopard now!!" This is a lot different than a true symbol for leopard, which can can be used counterfactually, with no direct signal content, so that I can say "do you think we'll see any leopards today?" and you aren't forced to respond "Leopard? Where? Run!"
Finally, about models: I think the idea of mental models in pre-linguistic species is interesting, but as I wrote to Billy, what would be their medium? What actual animal models can we point to, the way we can point to human ones? Speculation is interesting, and may turn out to be true, but does not warrant, I believe, such strong claims.
Posted by: Chris Schoen | Jun 30, 2009 7:54:54 PM
'Pre-linguistic species'... well put. Yes, it can be hard, even from so high up the scala natura, to understand how even those species way down below do their thinking, no? ;)
This question of yours neatly summarizes what I see as your impasse: "How can you have a representation without a symbolic structure (language) to do the representing?"
Good question!
Try to get past this (what I would call outdated) hangup on language and a whole new world of what thinking is will open up. I can represent an entire day of events, logic included, in my head without using a single word or linguistic formulation--is that just me or can't you do it too? There's more to representation than mere words. (Reporting those thoughts to others, though, often takes language--but not necessarily).
As to reasoning, how many of your decisions do you think are based on verbs and nouns? Don't you notice that your head goes 'quiet' when you actually have to think about something? How would one 'think' with words anyways? As I said above, they're much more useful for reporting (to oneself as well as others).
So, personally, I think language is much more about communication than cognition. In fact, I don't think our cognition has that much to do with language at all--but there are quite certainly those that disagree (I also think consciousness plays less of a role in reasoning than we, well, consciously, think it does).
I don't claim to know these answers to these questions, which are certainly huge and controversial--but things are not as clear-cut as you put them. I'm afraid, though, that I can only recommend dipping into one of the books I mentioned because I don't have the time to write more. As to consciousness, another nice source is:
Farber and Chuchland (1995), 'Consciousness and the neurosciences: philosophical and theoretical issues' in The Cognitive Neurosciences, MIT Press.
They are far more knowledgeable than me. Thanks for the interesting conversation, though!
Back to work...
Posted by: Kim | Jul 1, 2009 9:18:06 AM
Kim,
Thanks for taking time away from work to write about this. I realize you might not be able to devote any more time to this conversation, but I want you to know I remain deeply skeptical. If you can imagine a "new atheist" asking for corroboration of a deity, that's approximately how skeptical I am about concepts without symbols.
I want to observe here that you have changed some of your language. You write, for example, "how many of your decisions do you think are based on verbs and nouns?" But I never denied that rats or other creatures could make decisions (choices); simply that they could not make reasoned decisions (because this would require conceptualization of counterfactuals). And furthermore I do not mean to claim that because humans have a capacity for reason, that we base our decisions on it exclusively, or even in a majority of cases.
If you have a moment to briefly respond, I am curious to know what you mean when you say "I can represent an entire day of events, logic included, in my head without using a single word or linguistic formulation." I have no idea what you are describing here. But I'd really prefer to hear it from you, rather than going to the Churchlands, who tend to make me want to break things.
Posted by: Chris Schoen | Jul 1, 2009 11:20:31 AM
I'm not asking you to have faith, I'm asking you to be more skeptical of your conclusion by opening up to a wider range of possibilities.
What's the difference between a choice and a 'reasoned choice'? I think you would characterize 'choice' as non-cognitive and a 'reasoned choice' as cognitive--or at least translated into the terminology of my field. But, and I don't mean to do a bait-and-switch here, there's no doubt whatsoever that rats and other animals are cognitive. Behaviourism was killed and no one has regrets.
Language just provides the ability to explain, post hoc, why you decided on a certain course of action. But being able to articulate something to yourself or others doesn't mean that you must be able to articulate your reasoning in order to reason. I think most of our decision making works the same as for rats. We can talk about it, yes, but as I said before, the formulation of reasoning into language occurs, in my opinion, almost as an epiphenomenon (though not entirely: of course it can be used to guide future behaviour).
Representing a day... just imagine something. Imagine a scene in another country, imagine a situation, imagine how it would play out one way or another based on different imagined conditions. Man, just imagining something means you are mentally representing it! If I picture a squirrel in all its bushy-tailed glory, then I am mentally representing it. No linguistic grunt needs to be there. And I can reason my way through interactions with this squirrel, whether in my mind or actually squared off with the real thing in a field, without any additional grunts or networks of grunts. I do not see language as woven into things as intrinsically as you do. Don't believe all the hype!
Sorry, that might not be entirely coherent...
Posted by: Kim | Jul 1, 2009 12:05:48 PM
Kim,
Behaviorism is dead as a branch of human psychology. However I don't think the notion of stimulus-response is entirely unhelpful in biological descriptions. Whatever you may want to say about animal cognition, there is surely a threshold where no consciousness is understood to play a part: the turning of an oak leaf toward the sun, for example, is not a "reasoned choice," is it?
That sure *sounds* like you're asking me to take it on faith! If there's no doubt it should be fairly simple to explain why. Maybe there's actually a little bit of doubt, which is what makes this such a difficult topic?
If I understand you correctly you are arguing that images are the medium of representation in non-linguistic species like rats. I still find this unconvincing (especially paired with your equation of language to "grunts;" The fact that we can think silently doesn't mean we aren't using linguistic concepts.)
I think your argument could be turned the other way around, so that we could say that just because you aren't aware of the linguistic element of imagistic thought doesn't mean it isn't there. It could be operating on a pre-conscious or unconscious level. That doesn't seem any less plausible to me than arguing for rat cognition that there is no direct evidence for.
But I'm going to read and think about this a little longer before shooting my mouth off any further. If you send me your email address I can let you know when I have something more substantive to say. My name is linked to my blog, which has contact info.
Posted by: Chris Schoen | Jul 1, 2009 2:38:18 PM
Thanks for answering my questions, Chris. My reply may be obsolete now, but I'll give it anyway.
"I don't know how you could ever control for all of [the contingent factors] sufficiently to reliably correlate neural activity to mentation."
I think you suffer from a failure of imagination. Example 1, the experiment from Carlos's article. The scientists were able to determine, just from the activity in the rat's brain, where it would be in the maze and whether it was moving or not. Example 2, the fMRI and mind-reading experiments. The scientists were able to determine the object/orientation the person was thinking of. I think there's still an issue about what activities in the brain rise to the level of consciousness (although I haven't read anything about this for a while)--but that opens up another can of worms (for example, with split-brain patients, evidently both hemispheres are "conscious", but they can't talk to each other--so what exactly do we mean by "consciousness"?). In any event, I don't think the problem is as impossible as you think it is.
I don't know if the human/rat maze comparison experiment has ever been done, or what results to expect (i.e., is the brain activity fundamentally the same or not?). But I imagine that you're right, that humans may have a lot of unrelated thoughts, unless we're motivated to finish the maze as fast as possible. But even in that situation, I'd still say that we'd have "thoughts"--I guess this just means that we have different definitions of that word. Just an aside, about how we can have non-maze-related thoughts while running through a maze, maybe rats do too:
Memories may be formed throughout the day, not just while sleeping: http://www.physorg.com/news164376801.html
I came across an interesting paper, in one of the places Kim said to look. It describes what kind of experiments would or would not be evidence for an animal having a "theory of mind". I don't know if this type of thinking would fall under your category of "thoughts" or "symbolic mentation", but it's another example of how we can get at what's going on in minds without using language (not that asking people what/how they're thinking necessarily gives us an accurate picture of what/how they're thinking). The rat metacoginition experiment might not have been perfect, but it's also an example of this. It's hard to get at the mind without language or examining the brain, but not impossible. (Of course, I'm all for studying the brain to get at the mind.)
Posted by: billy | Jul 1, 2009 6:41:28 PM
Post a comment