Pavlov's plants

I decided that that thing about plants that make choices was too interesting not to follow-up on. Surely a plant that grows toward sunlight through a maze can't be said to be making a choice ? Growing towards sunlight is just what plants do, so there must be more to the experiment than that.

There is. I actually mentioned this in passing before, but it deserves a more thorough commentary.

Previously I suggested that in order to say something is really making a thoughtful choice, it has to be indirect : an entity must perform action A that causes process B to lead to consequence C. It can't just go straight to process B - it needs to do something that requires learned knowledge. Thinking about it more now, I suppose process B must also be context-dependent on action A. That is, if action A always leads to process B which in turn always leads to consequence C, then no choice is really being made : the consequence would still be direct, just several steps removed. Is that context-dependency enough to constitute a meaningful choice ? I don't know. Let's look at the plant experiment, maybe that will give some clues.

(I really do try quite hard to be self-consistent - see recent post about all the stuff I've got wrong - but the clue is in the name here : Decoherency. The whole point of this blog is to allow inconsistent ramblings and discussions that eventually lead to something useful, not to produce fully-formed ideas on the fly. So forgive me if I contradict what I said last time !)

Basically what they've done here is try and induce a Pavlovian response in plants. They start with a seedling at the branch of a simple Y-shaped "maze". On each fork they put a blue light, a fan, or both. They use a series of "training sessions" in which both the fan and light are present, in order that the plant might learn that fan=light and will therefore, in future, grow towards the fan based on the presence of the fan alone. They varied which arm of the maze the fan and light were placed in, so the plant's behaviour wouldn't be simply a directional preference. They trained the plants in three different ways : 1) with the fan and light on the same arm, so plants would associate a fan with light; 2) with a fan and light on different arms, so the plants would learn that fans do not provide light; 3) a control group with no fan or light, in case plants should have a natural directional preference for growth.

After training, all plants that were not stimulated by a fan or light grew towards the last direction of the light. But when they introduced a fan, something interesting happened. Of those in the group trained to associate the fan with light, 62% grew towards the fan without a light present. Perhaps more convincingly, of those that were trained to believe that fans do not provide light, 69% grew in the opposite direction to the fan.

This would seem to satisfy my criteria for a meaningful choice. Based on context, they take different actions in order to produce a consequence : if a fan is associated with light, they grow towards it; if it is associated with an absence of light, they grow away from it even with no light present anywhere else. I would have liked a another control training group with just a fan present, but that the plants will grow in the opposite direction may be sufficient by itself.

Let's assume the experimental result is valid (though I'll look up citations in a future post). Does that mean plants are thinking ?

I don't know. I find it hard to say they aren't making a choice, but then, the conditional if-then loops of computer code also constitute a choice. Is it fair to say computers are thinking ? Arguably yes, but in a completely mechanical, deterministic way. Of course, whether or not we think in completely determinstic ways is highly controversial by itself, but that's another story. For now : plants are capable of making choices. That's fascinating enough by itself.

Learning by Association in Plants

In complex and ever-changing environments, resources such as food are often scarce and unevenly distributed in space and time. Therefore, utilizing external cues to locate and remember high-quality sources allows more efficient foraging, thus increasing chances for survival. Associations between environmental cues and food are readily formed because of the tangible benefits they confer.