Ties between Center and Periphery in the Scientific World-System : Accumulation of Rewards , Dominance and Self-Reliance in the Center

This study account~ for the organization of scientific research in nctworb of socio-intellectual tics that bind scientists into a community cultivating the scientific tradition. During the twentieth century the scientific community ha~ become incr ca~ingly global both in the sense that its membership ha~ spread world-widely and in the sense that its long-distance tics have intensified. The globalization of the community and it~ tics ha~ been promot ed by widely institutionalized arrangements, especially through the world's adoption of and belief in several scientific tenets: the universal validity of scientific knowledge, the ownership principle that knowled ge should be the common property of humankind, and the political principle of granting autonom y to scientist~ for forming tics. The community and its network of tics form a hi erarchy with centers attract ing tics from peripheries. During the twentieth century the main center ha~ shifted from Western Europe to North America while Eastern Europe ha~ become less central, Ea~t A~ia has become a bit central, and other regions have remained peripheral. A center attracts students from around the world for education, attracts scientist~ for conferences and visit~, attracts deference from scientists throughout the world, exerts pcrva~ivc influenc e, is widely emulated, and is a desired source of recognition . In the global network~ of tics, specifica lly of deference, influence, emulation and desire for recognition, there is an accumulation in the center of tics, both from within the center a~ an enhanced self-reliance and from the periphery as an enhanced centrality, exceeding the research performance at the center.


Introduction
This study accounts for the organization of scientific research in a global network of tics binding scientists into a community, its institutional conditions, and its consequences for the communal cultivation of the scientific tradition, specifically for accumulation s of pcriphcrality and centrality in rewards and dominance.This scientific world-system has a communalit y illustrated by the following historical event: [Page 114]

Journa I of World-Systems Research
In 1900 in Paris, at the second "Congres International des Mathematiciens", a lecture on "Mathematische Probleme" was given by David Hilbert in which he outlined a number of new and old unsolved problems.Through their dissemination at the meeting and in its proceedings and journals around the world, they became known as "Die Hilbertschen Probleme".Hilbert's Problems became considered especially significant, and steps toward their solution would predictably earn much recognition; indeed, "a mathematician who had solved one of them thereby passed on to the honors class of the mathematical community", as a reviewer noted.Attention to these problems, their significance, and the recognition to be expected from contributing toward a solution were enhanced by their presentation by a person who was accomplished and central in the community.The dominance of a center was perceived by a foreign visitor to Hilbert's university in Gern1any (Bohr, 1947): • While UOttingcn was in 1nany ways a pcaccfill and quiet provincial tovv11, the richest scientific life Hourishcd there.An exceptionally intense and pure spirit of international brotherhood prevailed a111011g the young 1nathc1naticians, who fro111 nearly all countries went on a pilgri1nagc to the place, bound together by their con1111011 interest in and love iOr their scicncc ... UOttingcn was an international center for 1nathc1natics and other sciences ... nut over the whole lifC in UOttingcn, David Ililbcrt 1 s brilliant genius shone, as if binding us all together ... Ahnost every word he said, about problc1ns in our science and about things in general, sccn1cd to us strangely fresh and enriching, because everything bore the sta1np or his unique originality ... Thc whole present generation or 1nathc1naticians is -perhaps to a greater extent than 1nany or us ill.Hy realize-under the inilucncc or Ililbcrt 1 s work and the views he brought to triwnph.
The attraction to Giittingen and its wide-ranging influence illustrate the relationship between center and periphery.The center-periphery relationship was conceptualized by Shils in general terms (1988, pp. 251-252): • The ter111 nccntcrn rcfCrs to a sector or society [or conununity] in which certain activities which have special significance or functions arc relatively 1norc highly concentrated or 1norc intensively practiced than they arc in other parts or that society arrl which arc to a greater extent than arc other parts or society the fOcus or attention, preoccupation, obedience, dciCrcncc, or cn1ulation.
Propositions can be specified about centers; this paper will test the following hypothesis, exemplified by Hilbert's Giittingen, (Shils, [l 961] 1972, p. 357): • A quality, additional to the persuasiveness arising ronn the intrinsic value or idea or deed, gro\\rs silnply ifo111 cn1anation fro111 the 111ctropolis.The connection with the 111ctropolitan center coniCrs on an object or a syinbol a quality or its 0\\'11 quite indcpa:1dcntly or any inherent iCaturcs, so that 111uch or what co1ncs fro111 the center, even though it nlight be no better in itsclrthan what originates in the province, profits fro111 the special nature or its place or origin.
Such enhancement of the status ofa central participant illustrates the Matthew effect found among scientists by which recognition of research increases when performed by an already recognized scientist as opposed to a lower-standing scientist (Zuckern1an, 1989).Such inequitable allocation ofrewards is also found in studies of the unequal exchange in international trade with respect to capital accumulation in the world economy's center and impoverization in its periphery (Amin, 1974;Emmanuel, 1972).These lines of theoretical and empirical inquiry arc brought together here with a focus on tic accumulation in centers of science.
The hypothesis is not merely that tics accumulate in a center from around the world.It is posited that these tics accumulate not in proportion to the research performance of the center, but rather accrue in excess of predictions of its research performance.I shall t est the hypothesis that the center accumulates ties from throughout the community, both from within the center itself as an enhancement of its self-relianc e, and from peripheries a.., an enhancement of its centrality, over and above its research performance.
The accumulation in the center of tics from elsewhere around the world is an accumulation of its centrality.Scientists around the world arc extensively tied to peers in the center because of the central peers' high research performanc e. Howe ver, the hypothesis is that scientists around the world arc even more tied to their central peers than should be expected from the peers' research.This hypothesis can be formulated, equivalently, a.., the hypothesis that the peripheries arc relatively deprived or ignored , that a periphery attracts even fewer tics from around the world than expected from its research performance.

Journa I of World-Systems Research
The multiplication of a center's tics from within itself is an accumulation of self-reliance.Scientists there rely extensively on colleagues within the center because of these colleagues' high research performance.But the hypothesis is that scientists in the center rely even more on the colleagues than should be expected from th e colleagues' research.Because the center attracts so many tics from elsewhere, it is actually paradoxical and counterintuitive to posit that the center is especially non-outward.Such studies of world science have been characterized a.., geopolitical because of their focus on geopolitical groupings such a.., nations and regions (Shrum and Bankston, 1993/94) and a.., network analytic because of their modeling of tic networks within and between nations and regions (Knoke, 1990, pp. 20 1-202).More broadly, they arc studi es of a world-system and its globalization (Robertson 1992).In this study I first consider the twentieth-century global spread in the institutionalization of science a.., a tradition that is cultivated communally by scientists who form an incrca..,ingly global community with a network of tics that now span all regions .Then I map the global netwo rks of ties, specifically researc hers' deference toward places of scientific achic vcmcnt, their educational and scientific travel, their collaboration and influence, and finally their desire and competition for recognition a.., a reward .These networks reveal the centrality ve rsus pcriphcrality of eac h region and the outw ardness versus sc if-relianc e of the reg ion.Finally I test the hypothesis that ties accumula te in the center, both from around the world so a.., to amplify its centrali ty, and from within the center so a.., to deepen it.., self-reliance.I shall not exami ne the process of accumulation but its outco mes, the accumulated centrality and the accumulated self-reliance.But first I shall consider the institutionalization of communality a-; an institutional foundation for the formation of a hierarchy of centers and attached peripheries.

Institutionalization of Communality in Science
Science is a tradition that is cultivated through the performanc e of a social rol e that ha-; come to be called scientist.The script of the scientific role, the ta-;k of a scientist, is to cultivate science by contributing to the tradition, that is, by a-;similating the tradition, creating knowledge that is original juxtaposed to the tradition , and contributing this original knowledge to the tradition so it becomes a modified tradition.The scientist continually a-;similatcs the tradition by acquiring idea-; from other scientists, evaluates their validity and worth, utilizes some for further research, and in turn disseminates the results of the research to other scientists.In this enactment of the scientific role, th e scientist enters into role-relations with other scientists (Znanicck:i, [1940(Znanicck:i, [ ] 1968).The essence of the role-relationship is the exchange of idea-;.It is not only intellectual material that is exchanged.A contribution to the scientific tradition is rewarded with recognition, so knowledge is disseminated in exchange for recognition (Merton, 1973;Storer, 1966).A scientist's role-relations may thus include deferenc e to others and caring about being rewarded in form of recognition from other scientists.Furth ermor e, because recognition is awarded for originality and not rediscovery or triviality, a scientist competes with peers to be first and best, and emulation thereby enters into the scientist's role-relations with other scientists.The scientist thus enacts the scientific role in a circle of other scientists, and the role-relations with this circle arc not only intellectual exchanges but also social relations such a-; recognition and emulation (Grc, 1955 ;Znanicck:i, [1940Znanicck:i, [ ] 1968Znanicck:i, [ , 1965)).A scientist's tic with another scientist, a-; the tic actually exit-;, is a bundle of analytically distinct kinds of relations, som e of which arc int ellectual exchanges and some of which arc more social relations (Burt, 1985).Characterized in terms of specifici ty versus diffuseness a-; one of Parson's patt ern-variables, the tic is specific to the extent it only involves intellectual exchanges and it is diffuse to the extent it also involves social bonds .The social bond may not be cpiphcnomcnal to the intellectual exchange; rath er, there may be an cmbcddcdncss of the intellectual exchange in a social bond (Granovcttcr, 1985).The cultivation of science is thereby organized by the tics among scientist-;.The organization is neither that of a formalized hierarch y a-; in a bureaucrac y nor that of spot exchang es in a market, but is a network of tics that tend to form spontaneously, involve some trust, and acquire some stabilit y (Barber, 1987;Crane, 1972;Polan yi , 1967;Powell, 1990).Research is thus organized into collegial network-; that remain informal although they arc facilitated by a complex of int ernational organizations, ranging from international scientific and profes sional a-;sociation to UNESCO (Crane, 1981;Lyons 1963).

[Page 116] Journal of World-Systems Research
Through these role-relations and more social tics, scientists' cultivation of the scientific tradition is highly communal (Hagstrom, 1965).Scientists form a community, not only by sharing a received tradition, but also by their communal making of changes.Contributions arc subjected to rather communal evaluation and rewards arc allocated rather communally.Notably, the highest reward, the Nobel Prize, is awarded on the ba~is of evaluations from scientists around the world and thereby the reward is actually awarded on behalf of the world's community of scientists (Zuckerman, [1977(Zuckerman, [ ] 1996)).Scientists also form a community by their interpersonal tics.Scientists arc not merely fellow-members and peers but arc colleagues who form intellectual and social tics with one another.Their collegial tics bind them together into a community.A scientist defers to contributors, is influenced by the works of others, emulates others, and desires recognition from peers.These tics integrate the scientist into the community and form its ba~ic network.
The formation of tics, especially across long distances and social differences, depends on the institutionalization of science.The formation of a global scientific community ha~ been promoted by a worldwide institutionalization of the belief that validity of propositions is universal.Scientists arc convinced, partly by a~similation of a common tradition and partly by replication of experiments and observations, that truthfulness docs not depend on time or place (Shapin, 1994).The occa~ional loss of faith in universal validity, in the twentieth century mainly under Nazism and Soviet communism, ha~ on those occa~ions entailed some temporary disintegration in the community.
Furthermore, the formation of global tics also depends on the ownership principle that scientific knowledge should be the common and shared property of humankind (Da~ton, 1991).Conversely, appropriation of knowledge, whether by military secrecy or industrial patenting, counters the communal cultivation of the scientific tradition.
Communality also depends on autonomy.When science wa~ first institutionalized in England, the Royal Society of London in 1662 obtained its charter from the King who granted it autonomy to form tics throughout the world, "we have given and granted ... power and authority ... to enioy mutual intelligence and knowledge with all and all manner of'strangers and.foreigners" (translated from Latin and reprinted by the Royal Society of London, 1940, pp. 226-237).Since then, considerable autonomy ha~ been granted to scientist~ around the world to communicate with foreigners.Wars often constrain tics between scientists in nations that arc enemies.The Cold War between the communist Ea~tcrn Europe and the capitalist West hampered travels and interpersonal mail but scientific literature wa~ disseminated across the front, a~ scientists insisted on maintaining some autonomy and on continuing their communal cultivation of the scientific tradition, and Soviet scientists actually continued to exchange knowledge with, to defer to, to emulate, and to desire recognition from peers in the West (Schott, 1992b).Such a belief in universal validity, a common ownership of knowledge, and an autonomy to pursue tics with peers have become institutionalized around the world during the twentieth century.The institutionalization ha~ been promoted at the national and global levels, especially by organizations such a~ UNESCO, although autonomy is often constrained (Ziman ct al., 1986).These institutional arrangements arc a foundation for the formation of a hierarchy of centers and attached peripheries (Schott, 1991(Schott, , 1993(Schott, , 1995(Schott, , 1997(Schott, , 1998a(Schott, , 1998b)).

Centers of Achievement and Deference from Peripheries
The community of scientists is not a community of equals becaus e scientists di ffcr in their accomplishments, and its network is not a uniform grid.Indeed, an accomplished scientist attracts many tics while a novice is typically ignored.Tics arc especially dense between some participants and particularly sparse between some nodes.Tics arc dense within a country and sparse between different nations.Tics within and to a periphery arc sparse.The accomplishments of the center attract more tics, both from within the center and from peripheries.

Journa I of World-Systems Research
The center of science wa.., in Italy up to the seventeenth century when it shifted to England; then it shifted to France, then to Germany, and in the twentieth century the center shifted to the United States (Ben-David, [l97l] 1984).These nations wer e foci of attention because of their scientific achievements.The tics from scientists in periph eries to the centers have stimulated their creativity and have thereb y enhanced research in the peripheries (Schott, 1987).
The twentieth-century shifts of centers can be indicated by the research contributions that have become recognized by a Nobel Prize in science, awarded almost annually since 1901 for contributions in three field..,, chemistry, physics and medicine.This indicator of course falls short of perfection (in the pa..,t, apparently, the award.., by the Swedish Academy of Science favored contributions from the Nordic countries) but the distribution is actually very similar to the indication of deference obtained by a survey of scientists around the world (the later Table 9).The research recognized by a Nobel award is her e cla..,sificd according to its place and time of performance in Table l.
For this analysis, and especially for mapping the global network..., a cla..,sification is useful.Frequently used criteria for defining regions of the world arc either attributes such a.., geographical location, production and civilization, or clustering according to relations, such a.., cohesive cliques or positions given by similarity of patt erns of relati ons (Burt, 1983;Russett, 1967).Capturing the most common criteria, this study uses a cla..,sification of the world into the following eight regions: North Am erica, comprising the United States of America and Canada; Western Europe , comprising Germany, Franc e, Greece, Austria, Belgium, Denmark, Finland, Iceland, Ireland, Italy, Luxembourg, Netherland..,, Norway, Portug al, Spain, Sweden, Switzerland, and the United Kingdom; Israel, Australia and New Zealand; Ea..,tcrn Europe comprising the Soviet Union, Czechoslovakia, Albania, Bulgaria, Hungary, Poland, Romania, Yugoslavia, and their successors; Ea-.tern Asia comprising Japan, Hong Kong, Singapore, South Korea and Taiwan; Rest of Asia comprising Asia except Israel and Ea-.t crn Asia; Latin Am erica comprising South and Central America (including Mexico); and Africa.The only region that is not a geographical region is that oflsrael, Australia and New Zealand.It is rca-.onable to treat these a-. a region in this analysis insofar a-.they arc societi es established by European settlers who brought the scientific tradition with them.
[Page 118] Notes: An award shared between researchers is here counted fractionally among them.In each table the percentages in a column, before rounding, sum to 100% .

Journa I of World-Systems Research
Sources: Wasson (1987), and other biographies (listed in Crawford ct al. 1987).
Table 1 shows that, around the beginning of the 20th century, the main center of outstanding research wa<; in Germany.In the United States of America scientific activity wa<; surging and around the 1920s.American achievement reached a level similar to that in Germany.This finding is contrary to the rather popular image that it wa<; only after 1933 that the United States became the major center and that this centrality wa<; due mainly to the migration of scientists from Germany to the United States which followed the Nazi takeover in 1933.The center ha<; remained in the United States during the decades following the Second World War, with Western Europ e a<; the secondary center.
The survey is used extensively in this paper and can briefly be described a<; fol lows.Th e scientists were mostly sampled from the geographical index of th e authors of the articles in the journals covered by the Science Citation Index.The survey ha<; so far not covered Africa but analyses of the global networks of bibliographic citations and coauthorships show that the struc ture of ties of scientists in Africa is rather similar to the structures of tics of scientists in Latin America and Asia, so the lack of surveys in Africa docs not affect the generali zability of the conclusions of this study.My selection of a country within a region for surveying of course entails a reduction in representativeness, e.g.sampling Israeli scientists a<; representing scientist<; also in Australia and New Zealand.But, again, analyses of the global networks of coauthorships and citations show that the structure of ties from the sampled country is similar to the structur e of ties in the other countries in the region, and therefore we can be rca<;onably confident that the conclusions remain valid.The few earlier surveys comparing many countries, though, have focused on scientists' attributes and not on their tics (e.g.Franklin, 1988;Hcmptinne, 1990).
The scientist's deference to contributors wa.., tapped by the question "Who arc the five people in the world who have performed the best scientific research in your area of specialization since 1990?" (a.., worded in the American versi on used in 1995).The respondent named some contribut ors whose country wa.., also queried, thereby tapp ing deference within and among nations.
The network of deference from scientists toward contributors is mapped in Tabl e 2.
[Page 120] Journa I of World-Systems Research The first column in Table 2 shows that North American scientists defer mostly inward, to their local contributors, and little outward, to contributors in Western Europe and elsewhere around the world.The first row shows that North American contributors attract much deference from around the world.More generally, scientists in every region defer partly inward, toward local contributors, and partly outward, especially toward contributors in North America.
In every region the deference to contributors attracted from within the region exceeds the deference they attract from any other region.The high degree oflocal deference can be explained a-; a result partly of scientists' nationalistic pride in local accomplishment, and partly of their high awareness and familiarity with local work.
But in nearly every region the outward deference is larger than the local deference.The deference expressed by the scientists in one region compared to another is actually remarkably similar.The outward deference is not cosmopolitan but is directed largely toward contributors in North America and secondarily in West ern Europ e.
The region that attracts most deference and thereby is most central in the network of deference is evidently North America.The second-most central region in this network is Western Europe.The other regions arc peripheral in this network, they defer to the central regions but attract little deference.Juxtaposing the defer ences in Tabl e 2 with accomplishment earning Nobel award-; in Table 1, it is evident that a region's centrality or pcriphcrality in the network of deference reflects whether the region is a center or periphery of achievement.

Journa I of World-Systems Research
The deference entails an attraction that is expressed in travels.
People Carrying Ideas Around the World.
Ideas and knowledge are carried around by people when they move.Tacit knowl edge, uncodificd ideas and skills, cannot be writt en down for dissemination through publication, but is carried or embodied in people and is transmitted from its carri er to another person mainly through face-to-face interac tion (Polanyi, 1967).Mor eover, standards of evaluation, both for a-;certaining the importance of unsol ved probl ems and for judging the significance of recent contribu tions, arc not explicat ed, but evaluati ons arc of course ubiquitous and they tend to be formed through direct interaction among scientists.Persons pursue face-to-face interactions with others locally or they travel to interact with others.Travelers arc carriers of tacit knowledge and standards and, more specifically, transmitters of their idea~ and skills.A student who sojourns for research training in another place can return with research skills learnt from foreign teachers (Martin-Rovct, 1995).The researcher returning from foreign study is, in intellectual outlook, a stranger who is marginal to the local scientists, but it is the encounter of this imported learning with the local tradition which often results in a new combination of idea~.Likewise, a scientist who visits colleagues at another institution can acquire their tacit knowledge and use it upon return (Collins, 1974;MacKenzie and Spinardi, 1995).A visit may be short and involve only a few conversations and is thus a weak tic, but because the visitor comes from a different background than the host, each can present an idea that is new to the other and their idea~ may be combined into a new idea.The weak tic formed by a visit is thus important, a bridge between the somewhat different cultivation of science in the two places (Granovcttcr, 1983).Likewise, a scientist who migrates is a carrier of the locally created tacit knowledge that can be transmitted to and stimulate researchers in the country of immigration (Hoch, 1987).A piece of tacit knowledge is carried by a sojourner or migrant from its place of creation to another place where it can beget another piece of knowledge and a synthesis can be formed.Traveling thereby counteracts parochialism in the cultivation of the tradition of science (Hoch ct al., 1993).
The movement of research trainees and scicn tists is not a new phenomenon.Centuries ago, the intcrcivilizational encounter between China and Europe wa~ highly consequential.Ancient Greece and subsequently Europe had its wandering scholars and students, their traveling wa~ institutionalized, and when higher learning wa~ institutionalized in a corporation called a university, this corporation wa~ granted autonomy to move around.The institutionalization of education for young men in the European aristocracy included a Grand Tour around Europe.The nineteenth century German universities institutionalized Wanderiahre, the principle that education preferably included a year of study at another university, and in many countries advanced research training is considered complete only after a sojourn of training in the world center.In the twentieth century, the institutionalization of scientific travel ha~ been augmented by a global regime of rights to travel, right~ that arc not part of human right~ and thus not accorded to all humans, but professional right~ accorded to scientists for scientific purposes, although they arc often curtailed (Ziman ct al., 1986).Traveling ha~ become highly institutionalized, so much that its benefits arc taken for granted and policies of higher education and especially research training have in many countries around the world come to include policies for sojourning abroad (Goodwin, 1993).These policies arc not locally invented but arc formulated and promulgated by a complex of international organizations, mainly foundations and UNESCO with its constitutional mandate to promote 'the international exchange ofjJersons active in the.fields of' education, science and rnlture and the exchange ofjmblications' (the constitution is in, e.g., Bcstcrman, 1951, p.115).In peripheries of the world of learning, a sojourn to the center is a credential in itself, enhancing the prestige of the sojourner, and in some peripheral countries it is even somewhat of a necessary and sufficient condition for certain appointments (Goodman, 1984).In the center, conversely, foreign travel is a travel to a peripheral place, and is often considered worthless or even detrimental because fewer benefits arc thought to be obtainable by sojourning to the periphery than by remaining at home.This attitude prevails at the center, although its sojourners to more peripheral places a ... scrt that their sojourns arc highly beneficial (Martin-Rovct and Carlson, 1995).

Journa I of World-Systems Research
Traveling, especially travels to and from centers, ha.., promoted the global spread of science.The first major non-Western societies to institutionalize science were India and Japan and travels, both by local research trainees abroad and by foreign scientists to their societies, promoted the implantation and cultivation of scienc e in these societies.Around the beginning of the 20th century, thousands of Japanese and Indian students went to learn at universities in the West and hundreds of Western scientists sojourned to work in Japan and India (Burks, 1985;Shils, 1963;Singh, 1963).
Student"' travels around the world have been mapped by UNESCO since the middl e of the twentieth century.In 1950 the student.., enrolled in higher education abroad numb ered about .l million, the number grew over the years, and in the early 1990s the student.., abroad reached about 1.4 million (UNESCO, 1971 and1994), Table 3.The number of students abroad ha.., incrca..,cd much fa..,tcr than the population in the world.Indeed, the incrca..,c in foreign study ha.., been rather similar to expansion of higher education around the world (hence the percentage of the world's students who arc enroll ed abroad ha.., been rather constant, about two percent).Considering, mor eover, that opportuni ties for local higher education have grown around the world, so that the need to travel abroad for training has diminish ed, and considering the high costs of studying abroad, it is actually quit e remarkable that study abroad ha.:; been grow ing so fa.:;t.
Educational travel ha.:; not been random or uniform, but has crystallized along certain routes, Table 4.
[ Notes: The survey of host countries, reporting students from every nation, did not cover all countries (a~ also indicated in the table), so the number of students abroad from a nation is a little underestimated and the listed percentage in a host country is a little overestimated.
[Page 124] Journal of World-Systems Research Table 4 shows that students from around the world have traveled mainly to the United States, which since the Second World War ha.., been the center of higher education (Barnett and Wu, 1995;Ben-David, 1977;McMahon, 1992).Many students from th e former British colonies have studied in the United Kingdom, but this enrollment ha.., declined.Students from the communist countries sojourned mainl y to the Soviet Union, but recently they have mostly been going to the United States, a.., shown in the ca..,c of Czechoslovakia.Traveling between the Soviet-led communist Ea..,t European bloc and the capitalist W cstcrn bloc wa.., hampered by the rivalry between the blocs.Indeed, each bloc sought allies by attracting student.., from around the world.Apart from this cleavage, the flow in the twentieth century ha.., mainly been an attraction of peripheries to centers, mostly to Germany and the United Kingdom (especially from its colonies) in the first third of the century and since then mostly to the United States, the Unit ed Kin gdom, France, and also to the Soviet Union from the 1950s through the 1980s.Th e Soviet Union and the United States rivaled, a.., part of their Cold War, to attract students -and th ereby forge tics with the future leaders around the world -and with the end of the Cold War, efforts to attract students have diminished.In recent decades, the Europ ean Union ha.., promoted cohesion by institutionalizing and supporting movement of scienti sts and students among its member nations, so a rcgionaliz ation ha.., occurred in the Europ ean Union.But the European rcgionalization is not typical.On the whole, the distanc e that students travel ha.., been incrca..,ing, students' sojourns arc incr ca..,ingly to other worldrcgions rather than to other countries within their own region (Harmon, 1995).In terms of long-distanc e sojourns, there ha.., been a globali zation of higher education.
The educational bond between a teacher and a visiting student is likely to be succeeded, even upon the student's return home, by a professional tic between them.An analysis of the global networks of exchange agreements, student cxc hangcs, and collaboration and influenc e between scientists in different countries, shows that the exchan ge agreement.., between countries ha.., promot ed student exchanges between them, and, in turn, the student exchanges between the countries ha.., enhanced collaboration and influ ence between the countries' scientists (Schott, 1988).
Scientists, lik e student..,, have frequently sojourned to other countri es for intellectual intercourse (Martin-Rovct, 1995).Indeed, governments incrca..,ingly pursue policies of internationaliza tion of their national research enter prises .Occa..,ionally scientists mi grate permanently.Migra tion ha.., occurred in waves .With colonization some scientist.., sojourned and migrated from the colonizin g nations to the colonized country where they pursued research in the interest of the colonial power and facilitat ed the institut ionalizatio n of science in the colonized society and many expat riates stayed upon independence (Pycnson, 1985(Pycnson, , 1989(Pycnson, , 1993)).In societies established by European settlers some of them were also migrating scientist.-;.Notably, the cultivation of the scientific tradition in Israel was established with a wave of migrating scientist.-; in the int crwar period and ha.-; recently been intensified by a wave of migrating scientist.-;from th e fonncr Soviet Union (Kugel, 1993;Rodriguez, 1995) Most scientists in a country have been educated and trained locally, but some have also been educated abroad and arc either immigrants or repatriate s who have returned from studies abroad.The survey, described above, of sci entists in each region also a.-;k cd the scientists where they had received their highest degree.The educational origins of scientists is mapped in Table 5.The network includes scientists who immi grated to a region from another place a.-; well a.-; student.-; in a region who sojourned for hi gher education in another place and returned to become scientists.
[Page 125] Source: Survey (described earlier in the text).

Journal o.lWorld-Systems Research
The first column in Table 5 shows that North American scientists have mostly been educated locally, and a few have been educated in other regions.More generally, in nearly every region, most scientists have been educated locally, fewer have been educated elsewhere.In the surveyed Third World regions, though , around half of the scientists have been educated in other regions.The global netwo rk of education of scientists evidentl y ha-; two major centers, North America and Western Europe.
Formal education is typically a single event in a career and an educational tic declines in significance during the career.Another kind of tic may be more recurren t, that of traveling to conferences.The survey a-;kcd the scientists "To which scient[flc conferences abroad have you gone in the years 1990 to 1995?" (in the wording used in the American survey in 1995).The network of scientists' travel abroad to conferences is mapped in Table 6.
[  Table 7 shows that, in every region, scientists partly visit local institutions and partly visit institutions around the world.The global network of visiting, like the networks of education and conference travel~, evidently ha~ its centers in North America and Western Europe.Juxtaposing these various kinds of travels with the scientific achievements such a~ those earning the Nobel award (in the earlier Table 1) and deference to the achievement~ (in the earlier Table 2), it is evident that a region's centrality or peripherality in the networks of traveling mainly reflects whether the region is a center or periphery of achievement attracting deference.

Globalization of Collaboration.
Scientists travel to other institutions not only to disseminate their knowledge or acquire new knowledge, but also to confront their idea~ with colleagues' idea~ and thereby construct new knowledge, more or less a~ a synthesis of idea~ created in the different places.
Historical change, such a~ a process of globalization, cannot be tapped in a survey but can be examined by using archival sources, notably the scientific literature.Collaboration tends to result in publications that arc coauthored, so collaboration can be indicated by coauthorships.The publications authored by scientists in a country can be classified in three types, those coauthored with collaborators in other countri es, those coauthored between scientists at different institutions but all within the country, and those authored by scientists who arc all within a single institution in the country, Table 8.

[Page 128]
Journa I of World-Systems Research We can break the data on collaboration down further, Table 9.Source: Data ba~c acquired from Computer Horizons, Inc., derived from the Science Citation Index (cf.Stevens, 1990).
Table 9 shows that long-distance tics have grown even fa~tcr than tics to neighboring nations, thus indicating an incrca~c in the global span of collaboration, a globalization.

Circulation of Knowledge Around the World
Knowledge is carried around partly by people who travel and partly in other ways.Much scientific knowledge is disseminated through publications that, through the international postal regime, can and do circulate around the globe and arc a medium for the movement of knowledge.The flow of knowledge in the world scientifi c communit y can be crudely indicated by citations in scientists' publications referring to local and foreign authors (Zuckerman , 1987), Table 10.

[Page 130]
Journa I of World-Systems Research

Journal of World-Systems Researc h
Table 10 shows that knowledge flows extensively across long distances, transcending national and geopolitical boundaries.But the flow is not random and even.Researchers around the world assimilate and use knowledge flowing from centers, mainly from the United States.Even researchers in the Soviet Union utilized knowledge flowing from the United States almost as much as resea rchers elsewhere around the world.The flow evidently transcends national and geopolitical boundaries.
Whether a globalization of the movement of knowledge has occurred can be examined by longitudinal data on the flow of citations.Numbers arc available on the citations in articles published in 1976, and also citations in 1986, by scientists in the seven major countri es (USA, Canada, UK, France, West Germany, Japan, and the form er Soviet Union) to articles published in the most recent four years by authors in this set of countries ( citations in the Science Citation Index, similar to the data described in Stevens, 1990).The percentage of the citations within and among these nati ons which occurred between different nations was 34.05% in 1976 and the percentage increased to 34.79% in 1986.This increase in citations between, relative to within, countries, indicates a globalization of the movement of knowled ge.
The circulation of knowledge was tapped in the survey by asking the scientist "Who are the people whose ideas have influenced your research since about 1990? 11The respondent named up to twenty such influen ces whose country was also probed.The networ k of influence is mapped in Tab le 11.

Journa I of World-Systems Research
The first column in Table 11 shows that North American scientists receive influence mostly from peers in North America, some from Western Europe and little from elsewhere .The first row shows that North American science exert s pervasive influence on researchers in every region.1n every region, scientists arc partly inward, influenced by local colleagues, and partly outwa rd, influenced by peers ch,cwhcrc around the world.
Juxtaposing the map of influence in Table 11 with the map of deference in Table 2, it is evident that scientists arc more local in their received influence than in their deference.1n other words, the network of influence is somewhat less wide-ranging than the network of deference which extensively transcends boundaries to span the globe.The differences between the various networks will be examined later, in the section on outwardn ess.

Desire for Rewards and Emulation Around the World.
The institutionalization of a social activity commonl y includ es crystalli zation of a social role and a reward for performance of the role.1n science, the reward for scientific performance is recognition.lndeed, typically, a scientist cares about receiving recognition from peers and therefore also feels competition with others for recognition .
1n the survey, described earlier, a scientist's desire for recognition from others was tapped by a<;king, for each named influence, "To what extent do you care about each person's recognition ofyour research?"and the respondent rated the caring for recognition from each person on the scale from O for none, through 1 for little and 2 for some, up to 3 for great extent.This rated desire for recognition is used for calculating the distribution of the desire for recognition from various regions, Table 12.* Th e percentages in a column arc ba-;cd on the weighted distribution of influencers in which an influ encer is weighted by the respondent's rated extent of caring about recogn ition from the person (cf .. the earlier description of the survey).

Journa I of World-Systems Research
The first column in Table 12 shows that North American scientists desire recognition mostly from their local peers.The first row shows that North American scientists arc desired sources ofrccognition for scientists in every region.In every region, scientists desire recognition partly from local peers and partly from peers elsewhere around the world.
The network of desire for recognition transcends geopolitical boundaries like the tics of influence and only slightly less than the globe-spanning network of deference; this will become more evident in the later analysis of outwardness (Table 15).
Reward~ tend to be scarce and competition emerges.Scientists tend to emulate others in their performance of the scientific role.In the survey, described earlier, a scientist's emulation of others wa~ tapped by a~k:ing, for each named influence, "To what extent do you feel that the person and you are competing with one another to be.first or best in research?"and the respondent rated this emulation of each person on a scale from O for none, through 1 for little and 2 for some, up to 3 for great extent.The rated emulation is used for calculating the distribution of emulation across regions, Table 13.* The percentages in a column is ba..,ed on the weighted distribution of influenc ers in which an influencer is weighted by the respondent's rated extent of competition with th e person (cf .. the earlier description of the survey).

Journa I of World-Systems Research
The first column in Table 13 shows that North American scientists emulate local peers and occa..,ionally also peers in Western Europe.The first row shows that North Americans arc emulated by scientist.., in every region.In every region, scientists are partly inward, emulating local peers, and partly outward, emulating peers elsewh ere around the world.
Emulation transcends national and regional boundaries to the same degree a.., influ ence (mapped in the earlier Table 11) and only a little less than deference (mapped in the earlier Table 2).The similarities and differences among the networks will be examined later, in the section on outwardness (Table 15).

Accumulation of Rewards and Dominance in the Center.
The global networks reveal that the overall most central region is North America.To test the hypothesis that the centrality of the center exceeds it.., research, centrality shall be quantified.
The mappings introduced the terms outwardness and centrality to denote outward ties and attraction of tics, respectively.The centrality of a region in a network, more precisely, denotes it.., tendency to attract tics from elsewhere.Similarl y, the outwardness of a plac e denotes its tendency to send ties elsewhere.The ties from a place to another region are the ties that are sent outward from the place and arc attracted to the other reg ion.Therefore, the ties from the place to the other region depend on the tendency of th e plac e to send ties outward and also on the tendenc y of the other region to attract ties.Ind eed, the tics from the place to the other region may be proportional to the outwardness of the place and also proportional to the centrality of the other region.Hence, the tics from the place to the other reg ion may be model ed a.., the product of the two, Ties from a place to a region= Outwardness of the place * Centrality of the region.
Outwardness and centrality can be calculated from this model, using measur es of the tics from each area to each other area.Stat isticall y speaking, the measur ed ties arc represented in a two-wa y table where a column list.., the tics sent from a place and a row lists the tics attracted by a region, the local ties are disregarded, and the outwardness and centrality arc the para meters in the multiplicative model of qua..,i-independenc c of the rows and columns in the table, and these parameters can be estimated with common softwar e (e.g.Elia.., on, 1990, pp. 16-18).Th e outwardness param eter for a plac e depends essentially on how its column of ties to other area.., compare to the other columns of ties to the other area-., and does not depend directly on its local ties.Likewise, the centrality parameter for a region depends essentially on how its row of tics recei ved from elsewhere compares to the other rows of tics received from elsewhere.The centralit y of each region can thus be calculated from a table of mca-.uredties.
The centrality of a region in a network of tics is listed in Table 14, together with the regional share of research in the world.Research performance in a region is here indicated by its scientists' share of the articles published in the journals covered by the Science Citation Index.This indicator docs not merely tap volume but also incorporat es achievement because the Index mainly covers the journal-.that tend to publish considerable achievements.
[  ** Centrality of a region is given as its percentage of the sum total across all the regions.
Table 14 shows that the centrality of the main center, North America, substantially exceeds its research in six of the seven global networks (only in the network of conferences docs the measured centrality ofNorth America not exceed its measured research performance, but this is probably not substantively important but is likely an artifact of the comparative ovcrreprescntation of North American research in the data source, the Science Citation Index).Conversely, the tics to the periphery, namely Africa, Latin America and the Rest of Asia, in these six networks arc even fewer than predicted by the research in the periphery.This evidence corroborates the hypothesis that tics to the periphery are attenuated below its research and that tics to the center accumulate in excess of its research.
Research performance in a region was here operationalized by its share of publications in a North American reference work and this source may be somewhat biased, ovcrreprcscnting North American research and undcrreprescnting research performed in peripheries, and therefore the indicator may not be perfectly reliable.Notably, the North American research performance may be somewhat less than the indicat ed 39%.Fortunately, this unreliability does not affect the test.Specifically, if the North American performance is less than 39%, there would be even more reason to conclude that the North American centrality exceeds the performance.

Journa I of World-Systems Research
The enhanced dominance and rewards in the center is the accumulation of external ties to the center.The accumulation of internal tics within the center is the phenom enon of sclfrcliance .
Outwardness and the Accumulation of Self-Reliance in the Center.
The outwardness of a region denotes the tendency of its scientists to direct their tics toward peers outside the region.The outwardness of the region can be operationalized a-; the estimate of a parameter in the above model.The outwardn ess of each region in each network of interpersonal tics is listed in Table 15.(0utwardncss is not mca-;urcd in the three networks of traveling because that would be inappropriat e; specificall y, outwardness cannot validly be mca-;urcd from the network of education because this network includes migration; outwardness cannot validly be mca-;urcd from the network of traveling abroad to conferences because in one region traveling abroad may be most opportune outside the region whcrea-; in another region travel abroad may be most opportune within the region; and outwardness cannot be reliabl y mca-;urcd from the network of visiting institutions because in the survey in one region visits to institutions may be most opportune outside the region wherca-; in the survey in another region visits to institutions may be most opportune within the region). .5 Table 15 shows that outwardness differs among the regions, which will be compared shortly (Table 16), and differs among the networ k....The overall outwardness prevailing in a network, as indicated by the mean listed at the bottom of Table 15, is higher for deference than for the other nctwork'l.Scientists arc more outward deferential than they arc outward in their nctwork'l of influ ence, emulation and desir e for recogn ition.

[Page 137] Journal of World-Systems Research
The self-reliance of a region in a network of tics is the tendency of the scientists in the region to direct their tics toward peers within the region.The self-relianc e is thus the opposite of their outwardness and can therefore be mca'lurcd a'l the reciprocal of the outwardnes s (a'l noted above, this parame ter docs not directly reflect the local tics, so self-reliance is not a mere reflection oflocal tics but is a comparative characteristic).The self-reliance of each region in the four interpersonal networks is listed in Table 16.The first column in Table 16 shows that self-reliance in the deferential tics is far higher in North America than in any other region.The second column shows that North American research is also especially self-reliant in the network of susceptibility to influence from various regions.North American scientists arc also more locally reliant in their emulation of peers than scientists in any other region.The fourth column shows that North American scientists arc also especially locally reliant in their desire for recognition from peers.Across the networks, a~ indicated by the mean listed in the la~t column, North American research is considerably more self-reliant than the research in any other region.This evidence corroborates the hypothesis that the center is especially self-reliant.

Conclusions: Accumulation of Ties in the Global Community.
Up to the late-19th century science was institutionalized and communally cultivated within the W cstcrn civilization, as a world-system spanning Europe and later spanning the West, while every other civilization cultivated its indigenous traditions of knowledge.The 20th century historical process of globalization of science comprises two phenomena.First, science became global in that it spread and underwent institutionalization by obtaining legitimation, autonomy, support and organization in virtually every society around the world.Second, the cultivation of science became global in that it became cultivated communally as well as globally, through globespanning movements of ideas and people who by these globe -spanning collegial tics formed not only national communities but actually also a global community.This communality has been founded on widespread institutionalizations of a faith in universal validity of scientific knowledge, of a political economy of knowledge stipulating that scientific knowledge should belong to humankind, and of a principle of granting autonomy and even support to scientist.., to form collegial ties with peers around th e world.

Journa I of World-Systems Research
The movements of people, as examined in this article, for higher education, for training as scientist..,, for attending conferences abroad, and for visiting colleagues at other institutions, have been extended to span the globe.Their collaboration in the creation of new ideas ha.., become more global.The circulation of idea.., ha.., become mor e global.Scientists mostly defer to and desire recognition from distant colleagues.
The globalization of movem ent.., of people, idea.., and recognition did not entail any equalization but crystallized recurrently into a periphery and center form ation. Up to the early 1930s scientific achievement wa.., higher in Germany than anywhere c1..,c and made Germany the center and since the Second World War th e United States ha.., been the major center.Th e scientific achievement and idea.., attracted attention from peopl e throughout the world of learning, the temporary and permanent mi gration of people wa', mainly from the periphery to the center, and the circulation of idea.., wa.., mainl y from th e center to the periphery .
The recurrent unequal exchanges between periphery and center exacerbated the inequality between them.Attention to the periphery wa.., attenuat ed and its share becam e even less than its share of research.Conversely, centrality accumulated in the cent er and its shares of recognition and influence became even larger than its share of research.Th e accumulation of tics to the center from around the world wa.., accompa nied by an accumulation of tics from within itself.The center's scientist..,' deference, desir e for recognition and susceptibility to influence accumulated were even less outward than expected, so that the center wa.., especially self-reliant.
The tics from the peripheri es to the centers have stimulated research in the peripheries, elevated the dominanc e of the centers over and above its research perform ance, strengthened the cohes ion of the global scientific community, and enhanced th e coherence of the cultivation of the scientific tradition.
Science and its communal cultiva tion have become institutionaliz ed around the world, so much that its benefits, to individuals, to nations and to humankind , have become tak en for granted .Indeed, whcrca.., many states in the decades after th e Second World War pursued policies for the 'nationalization' ofresearch in the country, they have recently been elaborating policies of 'internationalization' of their national research enterprises, including policies for joint research facilities, support for collaboration among scientists, and for sending and receiving research trainees.The national policies have been not so much nationally invented a<; they have been formulated and promulgated by a web of international organizations led by agencies of modernity such a<; the Organisation for Economic Co-operation and Development (OECD) and Western-dominated agencies of modernization around the world such a<; the United Nations Educational, Scientific and Cultural Organization (UNESCO) with its constitutional mandate to promote 'the international exchange ofjJersons active in the.fields o.leducation, science and culture and the exchange o.fjmblications'.The globalization of science and its communal cultivation have thereby been part of the projects of modernization and globalization of social and intellectual life.

*
Res earch of a region is indicated by its percentage share of articl es by authors in the region, listed in the geographical section of the author index of the Science Citation Inde x from 1990 through 1994 (Philadelphi a: Institute for Scientific Informati on).

Table 1 .
Research for the Nobel award.

Table 2 .
Defer ence from scientists toward contributors.

Table 3 .
Students enrolled in higher education abroad.

Table 4 .
Students from each nation studying in other countries.

Table 7 .
Visits by scientists to other institutions.

Table 9 .
Articles authored within and among nations.Circulation of Knowledge Around the Worl d .

Table 11 .
Influence on scientists from peers.Desire for Rewards and Emulation Around the World .

Table 12 .
Recognition desired by scientists from colleague s.

Table 13 .
Emulation by scientists of colleagues.Accumulat ion of Rewards and Dominance in the Center .

Table 14 .
Research in each region and its centrality.

Table 1 .
Research for the Nobel award.

Table 6
shows that, from every region , scientists travel to conferences around th e world but the global netwo rk of conferencing ha<; its centers in N orth America and Western Europe .Another kind of tic that also involves travel ing is visits to other institutions.Visits to other institutio ns were tapped in the survey by a<;king "Which institutio ns have you visited in the last 12 months for purposes ofyour research?"The network of scientist<;' visits is mapped in Tab le 7.

Table 8 .
Articles by scientists in a country with coauthors in various countries; percent of the articles by the scientists in the country; 1973 and 1986.Table8shows that scientists around the world have collaborated rather extensively across national boundaries.Collaboration ha~ been centered on the United States.The United Kingdom ha~ been a secondary center.The Soviet Union wa~ a center within the communist Ea~tcrn Europe.Japan is becoming incrca~ingly central.As listed at the bottom of the table, scientist~ in a country arc dccrca~ingly pursuing their research within a single institution, but arc incrca~ingly collaborating, both with collaborators at other institutions within the country and with collaborators in other countries.Collaboration with collaborators at other institutions within the country, though, is not incrca~ing a~ fa~t a~ collaboration with foreign collaborators.The percentage of the coauthored articles that were coauthored between different nations wa~ 13% in 1973 and grew to 20% in 1986.

Table 11 .
Influence on scientists from peers.

Table 12 .
Recognition desired by scientists from colleagues.