To Karl Mannheim (7 December 1946)
 Dear Dr. Mannheim,
I have just finished reading “Diagnosis of Our Time” and felt that I would like to write and say how absorbing I found it.
Unfortunately I am not a sociologist, but I share for many years been extremely interested in education, and the study of education must of necessity involve a study of human society.
For some time past I have thinking on general terms about the problem of “How large should an organization be?, and your book has made me realise that this might be an important question in future planning. Attached are a few sketchy ideas upon this problem, which I feel has not received in the past the attention that it should.
I should be very interested, if you could spare the time, to know what your own views upon this problem are.
When Mr. Trevelyan, my Director, heard that I was writing, he asked me to convey his regards to you.
First notes upon the optimum size of an organization
 'Diagnosis of our Time' (Karl Mannheim) makes an incontrovertible case for viewing the society of the future as a planned organisation. It is immediately apparent that the optimum size for efficiency of an organisation should receive careful consideration. Too small an organisation may be inefficient in that it may have too much administration, or in that all possible resources are not completely utilised, while too large an organisation will fail because of the natural limitations to the powers of apprehension of human beings. The more complex a piece of machinery the greater the skill necessary to ensure that it works efficiency.
I do not know of any modern statement of this problem. The only relevant observations which I can recall are those in the "Republic" of Plato, referring to the bast size for a self-governing town. I am unable to verify Plato's figures, but I believe he put this down as 5,000 freemen and 30,000 slaves.
Examples of small organisation which have grown into larger ones. (i) The large number of relatively small railway companies which amalgamated into the larger companies and eventually became the "Big Five" as we know them today. (ii) The amalgamation of small private banks into a few large organisation is another case in point. (iii) The present-day tendency to research laboratories to be large affairs, rather than the individual and "backroom" kind typified by Cavendish or Dalton, would indicate that it as been found more efficient to do scientific upon team lines.
Examples of organisation which have been found too large to work efficiently (i) Ministry of Supply during the present war. This was divided into what were virtually watertight compartments, dealing with various aspects of the war supply problem.
(ii) The tendency with present-day large organisation to regard modern Management-theory as "from-the-bottom-up" rather than "from-the-top-down". See the article in the "Harvard Business Review", Summer, 1946, "Freedom within Management" by W.B. Given. This pleads for de-centralisation of a very large businesses. He quotes a case of a large company of 9 departments in which the financial procedure formerly was to assign so much money to each department. Today each of the 9 managers budgets for his own department and the President of the organisation acts almost wholely in a co-ordinating function. Managers of departments to their future as what they over-spend". Each manager is a creator of policy, rather than an administrator of policy which is dictated from above.
(iii) The Ministry of Education adheres to the policy of advising only on educational matters and not attempting to dictate to Local Education Authorities (usually a county). Is it possible that this is due to a felling that the county is the optimum administrative unit for educational matters?
If this problem is capable of solution, it seems obvious that the solution would differ for each organisation which is under consideration. No general answer can be given, and in any case empirical must be found.
I doubt very much whether the data necessary for an answer, except in the case of a few specialised organisations.
I do feel, however, that an explicit statement of the problem and some co-ordinating research into what has already been discovered would be of real value. What is needed is rather the correlation of existing date, than theoretical discussion of the problem. There can be little scope for predetermined experiment. Any experiment to be of value will need a nation-wide set-up. One is tempted to speculate upon whether the nationalisation of the mines and the proposed nationalisation of transport will appear to future historians as large-scale experiments upon this problem.
Possible sources of date: Russia must possess invaluable date bearing upon this question. Part of the solution might come from "Big-Business" which has devoted quite a fair of attention to this problem in recent years. Government experiments as the T.V.A. or the Boulder Dam work in America.
 Do some of the troubles of the Potsdam Agreement rise from the fact that the problems involved by Potsdam are too large completely to be apprehended by those who are supposed to solve them?
A pertinent quotation: -
"Not that I condemn intelligent planning, but at the moment we are wallowing in the interlude between laisse[z]-faire and a kind of statistical tyranny. The new age is being ushered in by a new type of young man, secretly envious, superficially clever, afraid of constructing ability, but obtaining a sense of power from energetic interference. They call it planning. At present it seems to me to resemble a vast paper factory. They do not understand life in the raw - man in the raw. They function through Acts of Parliament and bye-laws which no normal man understands.
That's the trouble with most administrators. They come to think of nothing but administration until there is nothing left to administrate, or they have fussed you into a strike".
From a recent article. Unfortunately the source has got lost, and I cannot trace the quotation which I found in a notebook I had kept during 1945.