Of Goldfish and Memory

This kid in Australia recently debunked the myth that goldfish only have 3-second memories. He did so by conditioning them to come to a beacon by associating it with food. He also trained them to follow short mazes using multiple beacons. These memories last for about a week, or indefinitely if used regularly. One of his goals was to show that it is cruel to put them in small aquariums/bowls. On a side note, it’s also cruel to put them in a small bowl with no air filter. They need at least one square foot of water surface per inch of goldfish (and multiple goldfish increase this proportionally). If you do less, you’re basically suffocating them slowly.

So, there has been a lot of work done on memory over the past hundred years. I have read very little about where modern theories of memory are, but I have thought about this a few times. It’s one of those things my mind wanders to while driving or otherwise engaged, and when I get back to some place I could check out more about it, I forget or don’t have time to do so. It’s interesting, just not pressing enough for me currently to devote time to. That said, here are some of my thoughts on the subject and maybe some random reader with more knowledge can shed some light or express an opinion.

It seems to me that there is a difference between things the mind learns about the nature of the world and things you catalog as memories. You know that knives are sharp and so don’t touch them. You were either told this as a kid or touched one and got cut. Probably both. You don’t remember when you learned it, but you did. You can walk, but you don’t remember how to do it or how you learned. You can add two numbers, but you aren’t counting them up in your head every time (usually). Maybe you remember learning this and maybe not, but you aren’t remembering the learning experience every time you do it. These strike me as low-level processes that your brain carries out quickly due to frequent use. If you were a computer, these things are in hardware. When you have a memory of Great-Aunt Sally pinching your cheeks, that’s in software. You consciously handle that memory and think about how uncomfortable that made you. When you remember your wife’s birthday, that’s in software (otherwise, she’ll take it out on your hardware).

So we have these memories we can manipulate with our conscious minds and then we have things we learned that are beneath the conscious level. Does conditioning an animal give them memories they access consciously and work on or does it imprint their hardware with some feature of the world? Can a dog remember when you utter some strange sequence of phonemes that you previously gave him a biscuit when you did so? Or does he just know (beneath the conscious level) that that sequence plus him sitting equals a treat? Now I’m pretty convinced dogs have more complex memories than that, but scale it down to a goldfish. Does this experiment show that they have conscious access to anything beyond 3 seconds? Or is it just conditioning? Is there a difference?

On the issue of dog memory, I know dogs can remember people (possibly either by sight or by smell) for at least a couple years. I also know they can remember the exact location of food for many hours. Since you might be thinking, well they just smell the food when they get around it again and so it’s not really remembering. My beagle has a nose like a champ, so he can sniff out just about any food in the same room as him. To watch him do it is amazing. But it takes time. He’ll lift his head and sniff the air, then start moving around and sniffing and then he hones in on it. Depending on the strength of the smell, it depends on just how fast he finds it. If he is pulled away from some dubious food item outside, he will rocket to it given the opportunity hours later. Even if the “food” has been removed entirely and the location is out of his line of sight when he is brought outside. No sniffing the wind. He just runs straight to it. Better experiments are perhaps required to be 100% sure, but the evidence looks convincing to me.