Re: testing memetics

Mark Mills (
Wed, 10 Dec 97 00:54:55 -0600

Message-Id: <>
Subject: Re: testing memetics
Date: Wed, 10 Dec 97 00:54:55 -0600
From: Mark Mills <>
To: "memetics list" <>

I recently heard about someone at the Viriginia State Aquarium =
training octopuses to 'open jars.' The success one has communicating =
the meme 'open jars' to octopuses is something that can be easily =
measured, so I thought I'd work with the idea a bit.

The trainer accomplished this in the following stages:

1. Open jars with shrimp or crab were set in the aquarium for the =
octopuses to find. This stage was complete when octopuses were =
comfortable locating and eating out of jars.

2. Lids were partially set over the food filled jars. This stage =
was complete when the octopuses quickly knocked off the lid to get to =
the food.

3. Lids were partially screwed down. The octopuses learned to =
translate their =8Cknock off the lid=B9 skill into unscrew the lid =
skills. Little by little, increase the =8Ctightness=B9 of the jar =
lid over a period of weeks. By the end of the process, the octopuses =
can open almost any closed jar placed in the aquarium.

Interestingly, the octopuses display a wide range of =8Cemotions=B9 =
while working on the jars. Octopuses are =8Cwhite=B9 when =
frightened, red when angry and flash between white and dark brown =
when excited. When they are frustrated, they pump water rapidly =
through a fluke in the back of the head. Trainers say they can read =
an =8Cemotional=B9 story into the octopus activity as they struggle =
with the jar, experiencing frustration and then excitement when they =
get the jar open. One is tempted to say they live out the same =
emotional stories that humans live.

There was no discussion of how long the octopuses could retain the =
skill (meme) without practice. As far as I could tell, the skill was =
not forgotten over nominal time periods like a week or month.

Apparently the experiment was something of a bet between aquarium =
managers. Some thought octopuses could be trained (communication =
established, meme transfered), others thought it impossible. As it =
turned out, the process of communicating the notion (meme) was very =
slow, but it happened.

What=B9s useful here is the simplicity of systems. The octopuses =
have a relatively simple neural system, probably several orders of =
magnitude fewer neural cells then humans. Additionally, they are =
relatively untainted with human memes which might confuse the =
measurements. Despite this limitation, octopuses seem full of =
memetic capability:

a) the ability to memorize behaviors
b) the ability to selectively use learned behaviors
c) expression of emotional states as part of the behavioral choice =

Octopus decision making syntax is probably much simpler than human. =
I don=B9t think this makes the behaviors non-memetic. The =
replication of jar opening behavior from human to octopus is memetic =
because the neural tissue of octopus records the necessary data and =
instructions for replicating the human behavior. The fact that the =
octopus is probably not going to teach other octopuses their jar =
tricks is a function of their social skills, not the memetic system. =
At some level, I'm sure you could identify some octopus social skills =
beyond the obvious sexual activities.

It is interesting to point out the lack of imitation. Memetics is =
not dependent on imitation in this case. The octopus never =
=8Cimitates=B9 a human. Octopuses are simply acting from their =
normal range of motivations and options. From this entirely octopus =
frame of reference, they are including memories of successful food =
gathering and including the key responses in their =8Cready=B9 =

The basis for their =8Cjar opening=B9 behavior is the ability to open =
clams and mollusks. Coupled with an in born desire to eat clams, =
they undoubtably inherit an ability to =8Clearn=B9 food gathering =
skills such as target selection and trapping techniques. In this =
case, all they need to learn is the shape and opening technique for =
this new kind of food source.

The ability to open a clam is an inherited meme. The clam opening =
meme is written into the neural system via some embryological =
development which included leaving =8Cblank=B9 part of the neural =
memory system for recording instructions for picking and =
=8Cpreparing=B9 desirable clams.

Memes are largely recorded in the neural tissue (substrate). If we =
knew how to physically read brain tissue, we would identify =
=8Cmemes=B9 much as people doing genetics identify =8Cgenes=B9 within =
the chromosome. As in genetics, one would have great difficulty =
drawing one-to-one causal relationships between 'bits' of the code, =
but one can identify what 'segment' of the substrate 'records' =
important parts of the organism's activity control system.


This was distributed via the memetics list associated with the
Journal of Memetics - Evolutionary Models of Information Transmission
For information about the journal and the list (e.g. unsubscribing)