Cephalopod Semantics

Ever since I was a boy I always wanted to keep an octopus as a pet. Believe it or not, I still do. And thanks to the power of the Internet, I know I’m not alone. Year after year I would passively research what I would need – three bio filters for the amount of bio waste it’s three tremendous hearts would output. The largest tank regular pet stores are likely to have (as the minimum!) and an interesting assortment of structures and movable objects. They are very very smart, and I would never consider taking on a pet if I knew it was in any kind of psychological distress or was emotionally of mentally needing for something.

And as I’ve grown, my expectations of what I might get out of this experience grew. Finally, an idea began to form that might just make it possible to teach an octopus to read! Sounds crazy doesn’t it?

Stage 1 – Reward Patience & Puzzle Solving Behaviour

Octopi are incredibly smart. One of the common problems facing owners is they are often smart enough to work out how to lift the lid of the tank! With enough time, they can work out how to unscrew jam jars by themselves.But what if we revived an old Pavlovian idea of conditioned learning, and fed the octopus by with food in the jar each time? Over time the Octopus should start to make the association. And Jam Jar would be king!

Stage 2 – Establish Visual Clues to Identify Reward

With this connection firmly made, what if we start introducing TWO brown smokey jars which were not transparent? One could have a ‘0′ for no reward. The other could have a marking to indicate prawn, both in different colours to help make them even more separable. If this process of feeding is continued, with enough time we can expect the Octopus to be more likely to go for the one it associates with the highest probability of getting food. This stage could take time, as part of a big learning curve but we will know when if/when it is complete because the octopus should stop going for the ‘zero’.

Stage 3 – Binding an Idea to a Word

Once this is mastered, a third jar could be introduced with a second type of prize – with a different symbol in a different colour on the jar. By this stage, the octopus should be fairly aware that this difference of label will indicate something about the contents and the feeding. It will be interesting to see how it reacts, which jar it goes for. Will it be driven by curiosity of the extra one? Will it go for the one it knows to be usually a reward, or will it go for the ‘zero’.

If it goes for the new one, we know that the experiment is working, and the octopus is curious, which indicates high level of intelligence. If it goes for the original symbol it was initially conditioned for, we know that it has worked, and the octopus is being driven by the slightly lower-level intelligence reasoning, based more around basic urges, coupled with an understanding of the likelihood of getting food. If it goes for the ‘zero’ we know the experiment itself is not working.

Perhaps two out of the three jars could be dropped each time it is fed? As we rotate through all combinations of the three jars, we could show the experiment is working, because hopefully it will avoid the ‘zero’ which will now act as the control.

It is important to say that the two possible treats on the menu now represent a choice. Even if that choice is only the resulting biological impulses stimulated by the proximity of the possibility of food. It cannot go for all three at once, and will need to decide.

Stage 4 – Abstracting the of Concept of Choice

If the octopus gets to the stage where we have the desired behaviour, flat 2-D laminated cards that look very similar to the original jars, with the same colours and markings could be introduced into the tank, far apart. Believe it or not, we share a very similar eye structure to the octopus, so we can expect our little icons to at least look identifiable.

The octopus could then be observed. Would it go for one of the reward ‘jars’? Will it investigate – led on by the primative drive in the belief that food might somehow happen? Especially with so much re-enforcement previously? If so, it can be rewarded with the appropriate choice, and hopefully this pattern might continue, with the symbolic choice now being made. The choosing mechanism has been abstracted.

The control ‘zero’ jar will now be very important, because the degree and frequency that it arouses the octopus’ interest, is the extent to which this experiment has failed.

Stage 5 – Abstracting the Choice Vocabulary

If the Octopus becomes comfortable with attacking or investigating the one it chooses, and with the pattern of being fed to help re-enforce this, we can try this one last phase. It might be that the abstraction stage was simply too much, and that we are now too far abstracted from the basic survival insticts for much of this to be visible to the creature.

BUT – if we know our pet well enough to identify it’s reaction and understand it’s behaviour – we can then just include three laminated square, flat cards that don’t look anything remotely like the jars, but still with the same symbols and colours.

In a successful test, we should see a clear decision by the octopus to investigate a particular card, and not being interested in the card with the ‘zero’. A failure, is that the octopus would show equal, or more interest with the ‘zero’ card.