SUCCESSIVE A P P R O S I M A T I O N IX UESCRIPTIYIC ECOLOO’i
41
for vegetation based on a variety of principles. Braun-Blanquet (1951)
has listed them conveniently as classifications based on physiognomy,
chorology, ecology, syngenetics (development), and floristic affinity.
There has been much argument among ecologists as t o which of these
most accurately represents the natural relationships of plant communities.
Gilmour (1951) and Cain (1959) have re-examined critically some of
the assumptions underlying classical plant and animal taxonomy, a i d
some of their comments and conclusions are very relevant. Gilmour,
especially, has examined the logical basis of classification and has made
a clear and invaluable summary of its aims and methods which I make
no apology for quoting in full.
“(i) Classification is a fundamental prerequisite of all conceptual
thought, whatever the subject matter of that thought. Thus, as soon as
we call an object ‘hot’, we are creating a class of hot things and
assigning the object t o that class.
“(ii) The primary function of classificatioii is to construct classes
about which we can make inductive generalizations. Thus, ha.ving
constructed a class of hot things, we construct a further class of ‘things
that hurt’, and, based on experience, we can make the inductive
generalization that ‘hot things hurt’.
“(iii) The particular classes we construct always arise in connection
with a particular purpose. The class of ‘things that hurt’ clearly
springs from a universally important purpose t o human beings - the
avoidance of suffering; the class of, say, ‘botanists whose fathers had
beards’, does not, and hence has not, so far as I know, been constructed
before this moment.
“(iv) The classification which we adopt for any set of objects depends
on the particular field in which we wish to makp, inductive generalizations. Different fields of generalization call for different classifications.
Thus, if we wish to investigate the connection between female authorship and murder stories, we would make quite a different classification
of books from that required for generalizations about, say, American
publishers and limited editions.
“(v) Clearly some classifications are of more general use than others.
Thus a classification of books based on subject matter. . . is of wider
use than one based on colours of bindings. The reason for this is that
the basis of the first classification, namely the contents of the book as
conceived by the author, affects more of its characteristics than does
the basis of the second classification, namely the colour of the book as
conceived by, presumably, the publisher. In short, subject matter has
a stronger influence on the total characteristics of books than has
binding design. In the general ttheory of classification, classifications
BZ
E.R.-I
41
for vegetation based on a variety of principles. Braun-Blanquet (1951)
has listed them conveniently as classifications based on physiognomy,
chorology, ecology, syngenetics (development), and floristic affinity.
There has been much argument among ecologists as t o which of these
most accurately represents the natural relationships of plant communities.
Gilmour (1951) and Cain (1959) have re-examined critically some of
the assumptions underlying classical plant and animal taxonomy, a i d
some of their comments and conclusions are very relevant. Gilmour,
especially, has examined the logical basis of classification and has made
a clear and invaluable summary of its aims and methods which I make
no apology for quoting in full.
“(i) Classification is a fundamental prerequisite of all conceptual
thought, whatever the subject matter of that thought. Thus, as soon as
we call an object ‘hot’, we are creating a class of hot things and
assigning the object t o that class.
“(ii) The primary function of classificatioii is to construct classes
about which we can make inductive generalizations. Thus, ha.ving
constructed a class of hot things, we construct a further class of ‘things
that hurt’, and, based on experience, we can make the inductive
generalization that ‘hot things hurt’.
“(iii) The particular classes we construct always arise in connection
with a particular purpose. The class of ‘things that hurt’ clearly
springs from a universally important purpose t o human beings - the
avoidance of suffering; the class of, say, ‘botanists whose fathers had
beards’, does not, and hence has not, so far as I know, been constructed
before this moment.
“(iv) The classification which we adopt for any set of objects depends
on the particular field in which we wish to makp, inductive generalizations. Different fields of generalization call for different classifications.
Thus, if we wish to investigate the connection between female authorship and murder stories, we would make quite a different classification
of books from that required for generalizations about, say, American
publishers and limited editions.
“(v) Clearly some classifications are of more general use than others.
Thus a classification of books based on subject matter. . . is of wider
use than one based on colours of bindings. The reason for this is that
the basis of the first classification, namely the contents of the book as
conceived by the author, affects more of its characteristics than does
the basis of the second classification, namely the colour of the book as
conceived by, presumably, the publisher. In short, subject matter has
a stronger influence on the total characteristics of books than has
binding design. In the general ttheory of classification, classifications
BZ
E.R.-I