Pavlov's plants are pants

Another follow-up post on the possibilities of plants that can think. Although there were some very vocal public criticisms, they didn't seem to amount to anything substantial, so I went and checked for citations to the original paper. Since I'm not a botanist, much less a philosopher, I resorted to using Google Scholar for this. Doubtless there are better ways to find citations, but this'll do. I skimmed through the paper titles, looking for any that might be relevant, then quickly checked the abstracts to see if that held true, and/or searching with ctrl+f for the author's name to see if the citations were actually important or just minor remarks.

There were 93 citations listed on Google Scholar. Of those, I found three which had anything of substance regarding Gagliano's claims that plants can learn and/or show Pavlovian responses to training.

The first has only a throwaway skeptical remark about the memory experiments, but some more interesting criticism about the Pavlovian tests. They say that contrary to what Gagliano states, that plants should normally grow in the direction of the last light source (the control case) is not at all expected, and that the review she cites does not describe this at all. What they say they expect to happen is that the plants should revert to vertical growth in the absence of light. If, then, there's a problem with Gagliano's control experiment, then the statistical significance of the training experiments is potentially greatly weakened.

The second paper is a review, so (quite rightly) doesn't do much besides state what Gagliano claims. However, it does mention that there might be an unexpected variable : plant-to-plant communication through chemical signalling. This means it's important to describe how the plants are stored, and the different stimuli the plants were exposed to could affect this. They suggest exposure to other subjects to control for this.

All this is quite interesting but doesn't actually undermine the central result of Gagliano's paper at all. However, as of April this year we now have a full-on replication study, and this shows exactly what the first paper was concerned about : the control conditions. The bottom line is that this study shows no evidence of any plant behaviour inconsistent with random chance. They try and recreate the original study as closely as possible, but it just doesn't seem to work. They do find that there is some tendency to grow towards the last light source, but it's much weaker than Gagliano's claim.

What's nice about this is that they emphasise the need for further testing with the original authors. Since their claim was statistically strong, presumably something interesting must be going on, but what ? They say that the author's co-operation would help address the (seemingly) minor differences between the two experiments, and they don't rule out the possibility that their own results might be in some way erroneous. So the evidence for Pavlovian responses in plants is, for now, looking a bit odd.