The Chicago lectures; Derek J. de Solla Price, part 2

Now is such unpublished work all that which is suppressed as a national secret not a contribution to science. I find in general that it's fair enough to say it is not. Silence is not silence. Communication lacks. In some ways it's a little hard to lose Cavendish from the ranks of contributing Santas and to admit new contribution. I know from the unpublished researches of Leonardo da Vinci. But on the whole it seems just a most serious but rather useful consequence is that this usage immediately distinguishes the science from a great deal of technology in which the end product may involve no publication. Oh only secret ones but only a new product or process or a post from the one which might even be even be recorded separately in a patent. Now definitional step then such as that which is published in scientific journals papers reports and books. Initial it is that which is embodied in the literature.

Conveniently enough this literature is far easier to define limit and count than anything else that one might deal with because of its central function. The sun just the literature has been subjected to centuries of systematize ation by indexes classifications journals of abstracts and retrieval for by far the greater part of the literature is now or has at one time been on the research front. We might distinguish this is research literature only LSA part consists of summer is another boiling stuff such as textbooks and popular treatise as all such literature can and in many cases has actually been counted classified and followed through the years as a time series for example the chief component of the research literature can be defined as that included in papers published in those Sunday big serial serials which are given in the list of scientific periodical a very

familiar told of all reference librarians. This list duplicate some journal titles. It obviously fails to include those particularly those near ofter its date of publication. It lists many general scholarly journals that include a few scientific papers but clearly it won't give all those that erupt with one scientific paper in a century. Well no With such a final step taken quantitative results begin to emerge very readily. We know to a first approximation that the number of journals concerned has been growing steadily since their inception in the 17th century. At a rate of 5 percent for atom compound interest. Now this gives you a doubling every 15 years a factor of 10 every 50 years and a factor of about 100000. Since the effective beginning about 17 we

know two journals have a very high mortality so that for every three that have been blown to have died and only one survives the commonest of journals is there for only about 35000 in the world. In fact we know rather more about this mortality every Jonah living. Every journal living about a hundred years. There are a hundred journals living in only one year every 50 living two years ten living and so on. We must overcome the difficulties of the statistical treatment as depended on a quite arbitrary selection of Judge rules by the world list. This can be converted to an objective method that tells much in the following form. We have now just beginning to be available of a new device called the silent citation. This takes about a thousand of the most important journals of the world in all fields of science takes every paper in them as it depicts this analyst spy off all the

papers and other sources cited by each of these original papers in its biographies and from now the proper use of this is that if you are interested in some field it will tell you what further material there is. Following papers that you or a bunch of colleagues or somebody else published last year. You see you can look up those papers in the index and see what new papers a citing this previous work. Thus there is a bit of graphical tool that can go forward in time rather than only backwards as in the old stop at the office thing. Now it's possible to use this day statistically for some purposes salting millions of facts about tape machine handling them. And quite apart from the practical use that this thing has the scientists the library. Even if we agree that much of the citation in papers perhaps most of it may be purely decorative for historical or personal reasons.

So it's plain that the statistical behavior in long which is regular and meaningful one can for example start from one journalist and tabulate the other journals that it cites and also the journals that cite it and these lists may then be set in rank order or total matrix can be established showing how each juggle fits with its neighbors. Note that this is largely independent of the original choice of Joe appear in the book. Even if say Physical Review had been totally emitted from the soul as soon as a few other journals publishing physics papers of being analyzed it would appear that Physical Review was full of them. A most important sectoral slight edge. That taking into account the number of papers available in each journal to provide citations on the number available in each journal to be cited. It appears still that some journalists clump together purely psyching

populations also in clumps. A fall from the collection being considered these clumps each consisting of a central coal surrounded by a few spotty unless a literature correspond roughly to the separate disciplines of so in principle I believe they can be mapped even on a plane with hybrid subjects like physical chemistry and chemical physics. Occupying the borderlands between the beginning as a physics and chemistry. Now not only could such a man be drawn quite objectively but one could also follow its path unchanging through the years so as to show the activity bursting out dying away there as fields come into being. Increase or decrease their production of papers or die completely. As an interesting health cover of such a mapping I think one finds quite a large class of allegedly scientific journals. But do not appear told or appear only on the pot on the far periphery of some of the clubs. That's to

say there appear to be journals that do not cite other journals much and I'm not into much cited. I would compare these functionally to newspapers each reflecting the world's events rather than consuming the products of all the newspapers and being consumed in that at to such a genus belongs quite a lot of the literature of the cloth is engineering or clinical medicine. All the data reporting now I don't quite understand the genesis of this growing but long cumulating literature but it seems to me some epiphenomenon of silence if silence it be a toll at all events one has now in principle a way of deciding objectively what metrosexual that is in the cumulation of sun and how it is divided by scientific sectors and grows at its various rights. By sampling a fair population of role 0

in any sector one can then determine statistically how this breaks down by countries by countries a publication that is by languages and we even have some trial investigations that show for the US in particular the source of funds supporting research published in paper as a way in which this change with time in the critical period when federal funds were beginning to replace private and universities a book without as it happens altering the growth rate of publication lump it in the fields considered of an interesting point now arises if we examine the breakdown by country and by field. It would appear as if most countries had amounts of the various sciences in proportion to the sizes of the sun tific fields and in proportion also to their own national scientific size. Now what's going on if there are found significant deviations from this total picture. They say physics is 15 percent of the total scientific get of the world.

But in the US it were 30 percent. Well I think there would be a good case for saying that this would be wrong policy and an economic unless some positive reason could be found or produced to the contrary. You see physics research is universal and an over contribution of one country would be adding more to the vote total than it could draw out as new developments of the research from and the money and the support with there will be better utilized in adding to the other relatively neglected side of this in proportion to listeners is on the world market. That's I think when I arrived at a situation where in principle the polis is for support a basic silence may be objectively handled and fundings suitably proportioned for those cases where one doesn't wish wilfully to disturb an evidence that is quo. Well having established an objective monitoring of the literature bun may now proceed to problems of scientific men here again the vital first step is one of

concepts and definitions for the same reasons as Biffle that of easing quantification. It's no strategic to define a scientist in terms of the literature. The easiest lied is to say that we should consider a person to be a scientist. If it any time in his life is all fluid. Oh perhaps even coauthored a paper in one of the scientific journals that we've already defined and classified in principle. Again there is little difficulty in showing that there is a mortality pattern in productivity which is relatively constant from country to country from field to field and even rather spectacularly constant with time so that it's ghastly change in two or three centuries. The natural distribution of productivity is and perhaps even degrees of goodness incompetence follow again something like a log no one would expect if the distribution of scientific talent was spread through the population. Like the heights of

men all their wealth or anything else. Well then there's the high mortality already found with journals which now tells us that for every three people once put to the research grant about to have been excrete it from it. And presumably into things that do not cumulate literature like teaching administration and the production of goods and processes as well as those that have been lost entirely into completely nonscientific activity. It can be seen that in principle if we know the volume of literature in any field in country in time we can compute the number of minimal scientists so defined and also the number of any higher degree of productivity that you might can tell me. In particular we can if we wish compute the number of call scientists who are responsible for the volume and more than half of the input book and we can also see the number that must be brought up to the point of publishing a single paper so as to maintain this population and enable it

to grow at the rate prescribed by the world total. Now I have a notion that there's a watertight no objective scheme whereby scientific literature and manpower may be measured and related one can expend the study in several directions. First one can take account of the logical place of known ideal definition an investigation that already has been made and relate them to this corpus thereby extending results in a very rigid practical way. Secondly one can consider the various statistical irregularities that have been found empirically and why they have the particular form observed in many cases we can now even begin to see certain underlying goals and principles that have a deep and obvious validity. It seems very likely that this sort of consideration will soon write its own areas of investigation from impaired a whole generalisation Kepler's laws to those which necessarily hold by virtue of basic

principles like the rules of new. Further goes the no well integrated Central corpus can be extended and deepened to link with historical and sociological studies on the one hand and with ramification in lines important to science policy discussions only help. I shall now it tempt to give a few selected examples for possible extensions of the central Colpus in these three different ways. Instances of the use of imperfect data even though we have no overall meant for the clumps of journals on a field by field basis. We have measurements for several of the most important drone also in various fields. But tell about the incidence of numbers of all the numbers of references from this data. It's quite clear that there is a field by field of variation a surprising aspect of multiple authorship is that it's remarkable growth

in some areas begun long before the incidence of big machines and the teams that are required to do experiments with them. And this leads one to suppose from the nature of the results that papers with 27 might well be regarded as a means of communication between the people involved. And the securing publication for people who would otherwise be a minimal producer and you know this the big machine is perhaps the occasion for the multiplicity rather than its goals. Further from the patent of the MLP by which papers contain citations it would appear that some areas notably in technologies and in clinical medicine feed on the previous literature very much less than is the goal. These subjects having only about 5 references but instead of the film about 15 Now does that mean

that in some subjects there is a different convention about how many papers one normally sign. I think not. It would appear to me that we have an effective difference indicating the function of the literature of these little citing fields is not primarily to provide data for the book because of the research from bio definition Munna rides at the provoking result that although literature is produced it is by this definition a patently weakly scientific in technology and in medicine. This instant says of the book in full of loads of love. Yeah that might be established in the silence of slow they're not but they might. Me I would like to consider that which tells us the nature of the distribution of size quality and Central and that which governs the normal way in which things measured grow with time. Now in all known instances this the distribution laws giving the number

of all those producing n papers o the number of institutions producing n son to us all that sort of thing one finds the same sort of pattern of extreme inequality. The fun has for the distribution of sizes of cities within a country or for the distribution of personal incomes in a highly competitive society. It's already been pointed out by Muslim sociological grounds that part of what is accounted for in silence by what he calls a Matthew principle and to him the task is given. To which one like well it use a complementary negative to him which has little is taken away completely. To put it in another way we have a situation similar to the effect of a bad law of Experimental Psychology in which each increase in effect is due to a constant increase in ratio of stimulus. The chunks you see the demand will go from his

first value but to his second which is about one challenge in full is the same as that of going from his second to his full o from his fall through his life of Miss hundreds to his to the chance of an institution growing from 10 Ph Ds in physics a year to 20 is the same as the chance of an institution going from 100 to 200 but gets there full in a normal population a random covariance distributing not the measures but rather their logarithms and on the whole this simple model provides a rather good first approximation. 0 0 approximation. Let me be humble to the empirical data. It's important to note that this is a built in mechanism of suck. So next is the sort of cybernetic control over its population it would be as difficult to make effective a decree establishing a level of the country a thousand equal institutions each producing 10 size Ph.D.s a year

as it would be to enforce a decision that will no huge cities be abolished and replaced by a population uniformly distributed over the map of the country. You counter with continuing this line can be seen that the same laws make it much easier to state the universally found exponential compound interest grows as a steady and linear growth. A uniform growth of the logarithm of all conventional measures of sun. Here again one can dig deep again using the technique of citation indexing in a current project I've been attempting to see what happens if one looks for the relation is not between one scientific journal and another but between individual office. It appears that the entire research front of Sauber may be divided into two distinct clumps of 0 each having a fountain a hundred maybe a little less members.

And within each clump the peer group are invisible colleges living by taking in each other's washing. The properties of these small clumps of particularly exciting for the social analysis. So there may be a significant same practical problems of administering and serving the needs in organizational literature for such groups. What's even more important is that we can begin to see that this is the mechanism for rapid accumulation of research through the interaction of a peer group which distinguishes the progress of science from that of scholarship nonscientific scholarship in general. In fact it's possible to derive a parameter which measures how much research in a field is growing from immediately previous book rather than drawing on the entire pulse of derived. Now this parameter then gives a measure of where that field of research stands in the

spectrum from pure science of learning to pure known style. Further one can see that it is in this short range interactions that make science grow so much faster than oneself and indeed one can derive a good approximate value for the growth rate in terms of a simple model of the interaction. It turns out from this that the old knowledge this brave new knowledge the constant right so that you have the differential equation giving you a linear growth of the lot of them as a necessary consequence of the basic situation rather than merely as an intent of finding. Lastly I must turn to a couple of samples of the way in which this new corpus of theory may be extended to meet the feelers being put out by those in urgent need of information and theory for practical planning. My first sample concerns the whole area of economic studies. So Bobo

is concerned with the financial cost of auto de and industrial growth. Don't you take me. The developed line gives a means of estimating field by field country by country. The number of people who are engaged at any time in producing the cumulating end product of literature that constitutes the stuff of scientific knowledge. It can also if we wish tell us on the same basis the effectiveness with which the accumulation is occurring at any part of the front or whether in detail on any specific group of researches is central in a clump or only peripheral. We do however have a measure of the man power directly involved and by multiplying by the average salary or income of such men. If you like sector by sector you can halve the amount of money spent directly on supporting the scientific population. And of course this can be broken down and followed with the years by the salsa nature of the funding or whatever else you wish. But something

fun can then add to this again second device thing. The amount that has been spent on a zillion resurfaces of technicians secretaries buildings Libres apparatus machinery and added together their results the total expenditure. It sounds a long rigmarole but it's relatively safe. You get the total expenditure on people silent together with the manpower directly and indirectly involved with the way in which the overall pattern agrees or disagrees with similar profiles of other countries in the world. Comparing these figures with national expenditures for research and development it then becomes obvious that in many sectors a great deal of money is being spent. That does not correspond to an output in cumulating scientific letters and is therefore not put just think that which we have defined as scientific knowledge. This by and large is the expenditure that is being made on technology on the production of new products and processes and services including medical

setting him on is no longer concerned with a universal and international state of knowledge to which each country subscribes but rather to a set of particular products that are being cut just by the country by the society for its military power prestige measures medical will being for industrial profit. Now the separation of this latter element from the US now enables one to say a little more clearly just what is being bought and how much is being paid for it. It now becomes possible for example to say how much one is paying full space son and how much for a lunar landing a communications satellite. How much in particle physics and how much in reactor technology. As against react to Suns perhaps of course there's some reciprocal relation between science and technology field by field but in the past we've never managed to estimate any dependence accurately because all previous figures of lumped

together on research and development in a somewhat confusing fashion. So my final of the example concerns deviations from the normal and stunted exponential growth of Son we know have several key examples that indicate that although the normal pattern is for doubling in size to occur every 15 years or even 10 new fields and new countries emerge exploding into a vacuum at a great to right field and overcome Trist act as if they're meeting with duration conditions that make growth considerably slow. A number of interesting case that is now exist to throw new light on the birth of a field. All the inception of science in a newly developing country. We know that overall the ill defined initial phase is the sudden onset of take off conditions and this is followed by a record growth to a certain point from which normal

exponential growth conditions are maintained. In a study of the growth of modern science in the late 19th century Japan it was shown that the crucial stage occurred with widespread reform of the educational system and a switch to teaching in Japanese as well as in European languages. Now such findings including the mathematical differential equations we have for the laws of motion of silence when you give it an initial impulse to move it. These things have direct applicability I think to the new problems of the developing nation. It is however in the possibility of an onset of saturation conditions in some countries notably in the United States and Soviet Union but none has perhaps the most provoking result of the sun. So it would seem that one of our most pressing problems is going to be that of the over developed countries where difficulties are already being

met not only in the supply of high talent man but in the rapidly increasing cost of utilising this. Because of this deceleration we would have the biggest scientific country is automatically taking a less and less a place in the world. And then they will be faced with even more difficult choices than we have now as to how to deploy their missiles and their town. So far as I can see the basic problems in organizing science and technology increasing very rapidly so that we're headed for demands this urgent as those which face the profession of economics during the financial crises of the Great Depression. It's with this demand in mind that I'm most concerned that we make every effort to have ready some well integrated and hard theory in the science of sun.

You heard Derek J Desoto price Avalon professor of the history of science at Yale University as he spoke on the topic. The science of science the subject for our next program will be brain mind and humanist values. Our speaker at that time will be Roger W. Sperry Hickson professor of psycho biology at the California Institute of Technology. This was another in a series of lectures given at the University of Chicago to initiate a new discussion on the nature of man. The lectures will be available in book form under the title on the nature of man to be published in late fall by the University of Chicago Press. These lectures were recorded by the Office of radio and television at the University of Chicago. The programs are prepared and distributed for broadcast by national educational radio. This is the national educational radio network.