Humans and the slime mould
One of the most general principles of human decision making is that we use relativity as a way to figure out how much we value things. We see a sale sign and the comparison of the current price to a more expensive past price makes us think that we are standing in front of a good deal. We see a modestly prices sweater next to a much more expensive one and we reason that is it a better deal for the money. And so on.
Relativity is not always the right strategy for figuring out how much to value things (very often it is not), but it gives us a quick and handy tool for going about the world making decisions.
Over the years the same type of relative decision making has been shown in monkeys, birds, and bees, but now it has been shown even with very simple lifeforms — the slime mould, Physarum polycephalum
Latty and Beekman did one such test using two food sources – one containing 3% oatmeal and covered in darkness (known as 3D), and another with 5% oatmeal that was brightly lit (5L). Bright light easily damages the slime mould, so it had to choose between a heftier but more irritating food source, and a smaller but more pleasant one. With no clear winner, it’s not surprising that the slime mould had no preference – it oozed towards each option just as often as the other.
But things changed when the researchers added a third option into the mix – a food source containing 1% oatmeal and shrouded in shadow (1D). This third alternative is clearly the inferior one, and the slime mould had little time for it. However, its presence changed the mould’s attitude toward the previous two options. Now, 80% of the slime mould headed towards the 3D source, while around 20% chose the brightly-lit 5L one. Even for slime mould relativity matters, suggesting that it is a very basic form of decision making!
The Honest Truth About Dishonesty: How We Lie to Everyone - Especially Ourselves
Very interesting.
The formal structure of this choice mechanism violates property beta, and not the independence of irrelevant alternatives as cited by the authors.
Pairwise indifference between 3d 5l, 1l is dominated on all pairwise comparisons but 3d stochastically dominates the choice set 3d, 5l, and 1l.
The reason the non stochastic version might be of interest to you is this.
Your thesis is that there a predictable but irrational choices. It would seem, however, that predictability implies something is being maximized – a difficulty for your thesis.
But this experiment provides you with a way out of this problem. Let me explain further.
Economists understand choice maximization in a very simply manner: Is there a preference relation that recovers the choice function, and is it fully transitive. Technically, we say that the relation R normalizes the choice function C, if R describes rational or maximizing behavior.
But this is a very interesting example of a) predictability of choice, but b) the underlying choice function not being normal – no maximization.
What you are interested in discovering is, then, the types of non-normal choice functions (irrational) that nonetheless describe predictable choice patterns. It would be especially nice to discover some simply conditions on non normal choice functions which describe reasonable versus unreasonable behavior.
Just because a behavior is predictable doesn’t mean that it maximizes some meaningful quantity. Even if a behavior maximizes something, it doesn’t count as rationality unless it specifically maximizes utility.
There is no preference relation that can be extracted from an observed choice function in cases where relativity is in play, slime choosing among various feeding choices being one example (Dr. Ariely details other examples in his books). The results of the “difficult to choose” situation show that agents (such as slime) do not prefer one option over the other (3D vs. 5L); yet when a third option is introduced, one of the two ORIGINAL options emerges as the preferable one. This means that a preference relation would have to say “option A and option B are equally preferable” in addition to saying “option A is preferred over option B”. There’s a really simple way to resolve this contradiction; the logical conclusion is that people do NOT have a preference relation that governs their behavior. Preference relations might be a good model in some situations, but relativity proves that it doesn’t always work.
Not all context dependent choices, non uniform choice functions, are going to be reasonable, or ones we would recommend.
So far we simply have observations or particulars about non uniform choice functions, we need some way to evaluate these choices. And a natural formal method is to expand the definition of maximization to include some but not all context dependent or non uniform choices. Otherwise we are left, because some technical results, that no context dependent choices are rational. That seems too strong, given the prevalence of context dependent choices which seem reasonable.
This comment is supposed to be in reply to michael’s second comment, but for some reason WordPress hates me, and won’t allow me to put this reply under his second comment.
Well, if a choice function is not even reasonable, it certainly cannot be construed as rational. At this point, the strongest thing you can say about behavior is that it’s predictable, which is exactly what Dr. Ariely is talking about.
As for a way to evaluate particular observed choice functions, the conventional wayto do this (as you point out) is to try to extract a preference relation; however, no preference relation exists that satisfies observed choice functions that demonstrate relativistic behavior. By this method of evaluation, humans behave irrationally, albeit predictably.
If you want a context sensitive explanation, relativity is exactly what you’re looking for.
It’s humbling to learn that a life form without a brain uses the same decision making “shortcut” that I catch myself using.
Wow! That’s amazing.
Wow that is amazing!
Howvewer I am on many dating sites and having met some slime mould, I can tell you I knew that long ago!
Looking at all the choices makes slime mould feel he can have anything and so writes to everyone, even the gorgeous ones they can’t have, but as soon as I show interest, mediocre turns out to be their sole mate. (no, I don’t spell it that way, they do…)
The irrationality of men who say they are open, honest and looking for love on date sites could garner you a plethora of data for classes and future books. maybe the guy who set up Plenty of fish will let you ask some questions there!
And Michael, what you say here is… ummm, well, let’s just say you need to stop over analyzing and trying to sound uber intelligent. it makes all the other mediocre stuff seem even more intelligent and I am drawn towards it!
“Otherwise we are left, because some technical results, that no context dependent choices are rational. That seems too strong, given the prevalence of context dependent choices which seem reasonable.”
We assume that such behavior is “guided” by (responds to) energy conservation principals. Not sure how it could exist otherwise.
The use of the word “choice” highlights the semantic dilemmas we face. What does the model/idea “choose” suggest/imply? Do mammals/primates have any different “choices” then slime molds? Probably not, we suspect.
Love you Book, by the way and I would love to chat with about it. Anyway, you ideas, in witch I believe, got me thing about what if we all were 100% racional? Everybodys talks about somekind of balance economic, but what if we were all rational. Would that balance stil stand? Or in fact waht we have now is rational world taking advantage on the irrational one?
Is this example not similar to the ‘The Economist’ subscription example you gave in your last book and the TED talk in 2008?
Am currently reading ‘Upside of Irrationality’ and am loving it. Thanks a lot.
I have a thought about how irrational we are when it comes to the environment. Would like to discuss this with you sometime if you have time. Thanks
Yes. How does this apply to our environment? And how does necessity and subjectivity play into things?
Staphanie, Human race in a whole seem to be pretty much oblivious to the environmental crisis (biodiversity loss, habitat loss, climate change etc.) that they are clearly contributing to. As rational individuals, we should have realized that harm done to the environment directly effect us as we are an integral part of an intricate web of life. It could have a direct bearing on our well being tomorrow and future generations. Thus we would have consciously reduced our negative impact for a better future.
But those of us trying to make the case for conservation and sustainable development often hit a wall`when convincing people and are often forced to follow unsustainable practices ourselves.
I see a long term irrationality versus a short term rationality in play here. The short term rationality being that “I maximize today’s well-being by all means”. But the presence of the longer term irrationality makes the whole process pretty much irrational don’t you think.
This is a reply to Karunakaran’s second comment, but for some reason wordpress is not giving me a reply link for that comment.
It seems quite likely that if we all behave rationally, we will all make decisions that end up destroying the environment. It may seem counter-intuitive, but this is a perfect example of what’s known as the Tragedy of the Commons. I recommend reading up on it in Wikipedia. The basic idea is that purely rational people will make eco-hostile choices, because the benefit one gains from leading an eco-friendly life is less than the value of the comfort and convenience that they’d be forced to give up. It’s true that if everyone made eco-friendly choices, the Earth would be in better condition, but no one is deciding what the everyone else should do; a person can only decide what he himself will do. The upshot is that we’re actually better off if everyone does the irrational thing and leads eco-friendly lives. I guess this is another one of those upsides of irrationality that Dan alludes to in the title of his latest book.
@Allcode;
No, there are a number of non-normal choice functions will look both reasonable and rational.
Think of any standard decision method which uses decision criteria to eliminate choices, starting with the entire group. For example, suppose we wanted purchase a house and our criteria were: price, school district, length of time to downtown, and size of backyard. Apply these in a lexical manner.
This choice function will not be generated or recovered from the choice behavior on pairwise comparisons.
The easiest way to see this is that all houses over a certain price are equally bad, being rejected. But on a pairwise comparison, between two different house prices, one of these equally bad choices will dominate.
Really interesting! Coming from an engineering background, I’d like to model this experiment with a simple mathematical one and get to the root of this.
Let’s say you are given two random unknown numbers a and b and you are asked to choose the larger one. Well since the numbers are random you would hesitate and choose either a or b.
Now assume you are given a third random number c, and you are told that b is larger than c. Again you are asked to choose a or b.
Now a rational thinker would say that the fact that ” b > c ” has nothing to do with the value of a. So he would disregard this additional information and get back to the original problem of ” a ? b”.
But if you let primitive emotions decide, the fact that ” b >c”, gives a certain sense of superiority to b in our mind. You would say that after all b is larger than c, and it could be that it is also larger than a. I mean I have no reason to believe that a is larger than b. So I’ll stick to b.
The reason for such a decision I think is that our brain yearns for structure and for being determined. Randomness seems scary and at least not desirable. So somehow we intentionally overvalue the information “b>c” in our mind so that we can get out of this state of randomness and indecisiveness.
Of course these are all speculations. Thanks for sharing this interesting study.