LEARNING BY MARINE INVERTEBRATES
51
ing the pedal disc and swimming away when touched by the starfish
Dermusterias (Yentsch and Pierce, 1955). In training Stomphia escape
was prevented by prodding the anemone, causing it to close up at the
same time as contact was made with the starfish. After ten to twenty
such trials the normal response to Dermmterim was lost (Ross, 1966).
How many of the coelenterate results one regards as indicating
learning depends, as pointed out at the beginning, on one’s definitions.
And thereby raises the more fundamental problem inherent in attempts
to define learning anyway. The fact of the matter is that at present
we label as “ learning ’’ all individual adaptive phenomena for which
we have no adequate physiological explanation. Where it is possible to
give an explanation, as for example in Jennings’ (1905) analysis of
position habit learning in Actinia, the phenomenon is struck off the
“ learning ” list. Sensory adaptation or sensitization are further types
of explanation that, when proven, withdraw the phenomenon from the
list. In effect this means defining learning in terms of central nervous
activity. We do not know the mechanism of adaptive changes in the
CNS and, indeed, have no good reason a priori for thinking in terms of
one mechanism only. All animals, so far as we know, have nerve cells
with rather similar properties. There are a number of theoretical
ways of connecting these up to provide circuits capable of holding a
record, even supposing that one is considering only explanations in
terms of synaptic connexions. When explanations in terms of local
dendritic potentials or chemical coding are added the range of possibilities is still further increased, and it may well be that different
groups of animals have exploited different nerve cell characteristics.
The requirements of a starfish are very different from those of an
octopus. In evolving towards their respective ways of life, echinoderms and cephalopods may have advantageously developed two
quite different computer systems and it would be remarkable to find
that both had exploited the properties of nerve cells in exactly the
same way.
IX. DISCUSSION
This section would perhaps be better entitled “ concluding remarks
and apologia”. When I agreed to write this account, I had in mind a
critical review of formal training experiments, with careful examination
of experimental techniques. It very soon became obvious that this
would produce a very one-sided view of the situation and in any caae
lead to the omission of much of the very matter from which the most
important criticisms arise. So I have tried to cover not only formal
training experiments but also other evidence of individual adaptive
phenomena-a
much more heterogeneous collection of material.
51
ing the pedal disc and swimming away when touched by the starfish
Dermusterias (Yentsch and Pierce, 1955). In training Stomphia escape
was prevented by prodding the anemone, causing it to close up at the
same time as contact was made with the starfish. After ten to twenty
such trials the normal response to Dermmterim was lost (Ross, 1966).
How many of the coelenterate results one regards as indicating
learning depends, as pointed out at the beginning, on one’s definitions.
And thereby raises the more fundamental problem inherent in attempts
to define learning anyway. The fact of the matter is that at present
we label as “ learning ’’ all individual adaptive phenomena for which
we have no adequate physiological explanation. Where it is possible to
give an explanation, as for example in Jennings’ (1905) analysis of
position habit learning in Actinia, the phenomenon is struck off the
“ learning ” list. Sensory adaptation or sensitization are further types
of explanation that, when proven, withdraw the phenomenon from the
list. In effect this means defining learning in terms of central nervous
activity. We do not know the mechanism of adaptive changes in the
CNS and, indeed, have no good reason a priori for thinking in terms of
one mechanism only. All animals, so far as we know, have nerve cells
with rather similar properties. There are a number of theoretical
ways of connecting these up to provide circuits capable of holding a
record, even supposing that one is considering only explanations in
terms of synaptic connexions. When explanations in terms of local
dendritic potentials or chemical coding are added the range of possibilities is still further increased, and it may well be that different
groups of animals have exploited different nerve cell characteristics.
The requirements of a starfish are very different from those of an
octopus. In evolving towards their respective ways of life, echinoderms and cephalopods may have advantageously developed two
quite different computer systems and it would be remarkable to find
that both had exploited the properties of nerve cells in exactly the
same way.
IX. DISCUSSION
This section would perhaps be better entitled “ concluding remarks
and apologia”. When I agreed to write this account, I had in mind a
critical review of formal training experiments, with careful examination
of experimental techniques. It very soon became obvious that this
would produce a very one-sided view of the situation and in any caae
lead to the omission of much of the very matter from which the most
important criticisms arise. So I have tried to cover not only formal
training experiments but also other evidence of individual adaptive
phenomena-a
much more heterogeneous collection of material.