HERBERT SIMON, PAUL THAGARD, PAT LANGLEY AND OTHERS ON DISCOVERY SYSTEMS
APPENDIX I- Page 14
Laumann’s second criticism and Hickey’s rejoinders:
Laumann: Laumann’s second chosen referee starts by writing that he has a great deal of sympathy with attempts to provide large-scale explanations of change in American society. But he next demands “concrete” thinking about “specific” mechanisms that are sources of stability and change. He says that “clumsy abstractions” such as Hickey’s latent control structures or Land’s opportunity-structures approach are “too vague” to illuminate anything about social change that is “not obvious”. He says that empirical analysis needs “disciplined” and “substantively informed” investigations to evaluate conjectures about social change, and he addresses Hickey saying: “In this regard I am simply not convinced by anything you report.”
He continues by saying that Hickey “eschews substantive reasoning” as a basis for the empirical models and instead let his computer program alone determine the specifications that he reports. He rejects the findings as often “bizarre”, as exemplified by birth rate varying directly with the homicide rate and inversely with expenditures on mass media, and says he finds this impossible to take seriously. He claims there is a burgeoning literature reporting analyses of the various influences of trends of women’s labor force participation rate, relative economic status, past demographic swings, and sex rôle attitudes on fertility trends. He says that these analyses in contrast to Hickey’s rest on attempts to specify “concrete behavioral mechanisms” responsible for observed trends, while Hickey’s equations are post hoc interpretations, which are never buttressed by independent evidence and hence are highly fanciful.
Hickey: Phrases like “disciplined” and “substantively informed” are romantic rhetoric. The Encyclopedia of the Social Sciences defines functionalism as an explanation of how major social patterns operate to maintain the integration or adaptation of larger social systems. In Hickey’s model the larger system is the national macrosociety. More formally stated functionalist explanations are about movements within a system toward stable self-maintaining equilibria. Functionalism lends itself to the distinctively macrolevel perspective. Durkheim saw macrosociety as an entity sui generis in need of social integration, which is not explainable by reduction to individuals or their psychological dispositions or motivations however “specific” or “concrete” they may be described. The anthropologist Malinowski stressed the integrating functions of different types of institutions for maintaining social structure. Hickey agrees with the insights of these authors, and he follows in the empirical emphasis on outcomes advocated by Merton. But Hickey substitutes mathematical modeling of the social system for the traditional analogies like homeostasis within biological organisms for describing equilibrium-seeking movements. His empirical model shows that a stable equilibrium does not exist in the American macrosociety. The “concrete” and “specific” mechanisms that are sources of stability and change are revealed by the simulations made with the model’s negative feedback relations.
Laumann’s second chosen referee needs no introduction for economists. His fallacious criticism is familiar and has long ago been repudiated. It is the same kind of dogmatic romanticism exemplified by Joseph A. Schumpeter, a member of the classical Austrian school of economics. In his review of Keynes’ General Theory of Employment, Interest, and Money in the Journal of the American Statistical Association (Dec. 1936) Schumpeter described Keynes’ propensity to consume thesis as nothing but a deus ex machina that is valueless, if we do not understand the “mechanism”. The economic historian Mark Blaug of the University of London writes in his Economic History and the History of Economics that Keynes’ consumption function is not derived from individual maximizing behavior, but is instead a bold inference based on the known relationship between aggregate consumer expenditures and aggregate national income (P. 243). Harvard’s Alvin Hansen called Keynes’ consumption function his greatest contribution. Schumpeter goes on to write that Keynes’ inducement to invest, his multiplier, and his liquidity preference theses, are all “an Olympus” of such hypotheses which should be replaced by concepts drawn from the economic processes and mechanisms that lie behind the surface phenomena. By “economic processes” and “mechanism” he meant the ersatz maximizing psychology of classical economics. Were it not for the Great Depression, the economics profession might have joined reactionaries such as Schumpeter in denying recognition of a distinctive macroeconomic perspective, as Laumann and his chosen referees are attempting to deny recognition of a distinctive macrosociological perspective.
Laumann’s second chosen referee says that he has a great deal of sympathy with attempts to provide “large-scale” explanations of change in American society. In his book Social Causation Robert (P. 371) Robert MacIver had expressed the view of classical sociology by saying that the primary contrast between social causation and the causation revealed in physical and in biological phenomena is that the former involves the “socio-psychological nexus”. But macrosociology is not just “large-scale” social psychology. The word “macro” means that the macrosociological theory describes phenomena that cannot successfully be reduced to microlevel “mechanisms” without committing the logical fallacy of composition. Unlike the old classical economists and Laumann’s classical sociologists, undergraduate students of economics today recognize the fallacy of composition, which sociologists can find explained at length in the introductory textbook, Economics, (P. 14) written by 1970 Nobel-laureate Keynesian economist Paul Samuelson. In Parsons’ terms but contrary to Parsons, macrosociological outcomes are “emergentist”, and this precludes reduction to an individualistic microlevel social psychology of motivational “mechanisms”. The motivations of social members cannot explain unintended outcomes exhibited by the macrosocial system. Explanation of unintended outcomes requires a macrosociometric model.
Laumann’s second chosen referee is no less dogmatically dismissive than were the 1930’s classical economists such as Schumpeter. Laumann’s second chosen referee is just more retarded, since Hickey’s paper was submitted to this sociology journal a half-century after Keynes. In the urgency of the Great Depression economists became pragmatic, simply ignored Schumpeter’s classical criticisms, and instead explored, developed and applied Keynesian macroeconomics. And in the 1950’s, when computers became available to universities, economists developed macroeconometric models such as the Klein-Goldberger model based on Keynes’ insights, which is how Keynes is still used today.
Sociologists should likewise ignore reductionist demands for “concrete” thinking about “specific” social-psychological “mechanisms”. They should not be deterred from developing empirical macrosociometric theories having a distinctively macro perspective. They should not capitulate to any editor or his favorite referee attempting to impose a dogmatic social-psychological reductionist fetish, as Schumpeter and his fellow Austrian-school reactionaries had attempted to bully economists with their economic-psychological reductionist fetish. Nor should Laumann’s Luddite referee bully sociologists for using computerized discovery systems to explore data in order to develop empirical models pragmatically, models that exhibit relations among the institutional variables that social-psychological reductionism cannot reveal. For example the integrative mechanism exhibited in Simulation II produces a dampening negative feedback due to equations (3) and (4), which determine the change rate of the compliance rate (LW) and the change rate of the high-school completion rate (HS). But this feedback is not due to high-school graduates’ motivations to produce macrosocial stability by increasing compliance with laws prohibiting homicide; it is an unintended outcome and is not explicable in terms of the motivations of the participants. Similarly the disintegration of the institutional order and the social disorganization in the macrosociety due to accelerated massive internal migration into cities is unintended and is not the motivation of the migrants.
Laumann’s chosen referee’s phrase “clumsy abstractions” is a clumsy dismissal. He writes that the clumsy abstractions are too vague to illuminate anything about social change that is not obvious. But if the findings from Hickey’s model are obvious, then why does this referee write that he is not convinced by anything Hickey reports? In fact he says this because the patterns and outcomes revealed by the model’s structure and simulations are not obvious without the model. Hickey did not invoke Ogburn or Land’s thesis of “opportunity structures”. But Hickey did invoke the Columbia University sociologist Robert K. Merton’s thesis of latent control structures, which may be operative where the relationships described by the equations are not obvious to the social members and to sociologists such as this referee. One of the most difficult problems with romantic attempts to reference motivations is that the outcomes of many actions are not explicable in terms of conscious motives, i.e., “concrete” and “specific” mechanisms, when the outcomes are unforeseen by the social participants.
Like the Institutionalist economist Wesley C. Mitchell, the sociologist Robert K. Merton said in his Social Theory and Social Structure that the concept of social function should refer to observable objective consequences and not to covert subjective dispositions such as aims, motives, or purposes that have enthralled the romantics. Observable objective consequences are what the model’s simulations describe. Merton makes the concept of function involve the standpoint of the observer and not necessarily that of the participant. He adds that the social consequences of sociological interest are those for the larger structures in which the functions are contained. This is a central thesis of functionalism, and it is central to Hickey’s paper where the larger structure of interest is the U.S. national macrosociety. Merton wisely warned that failure to distinguish between the objective sociological consequence and the subjective disposition inevitably leads to confusion, because the subjective disposition may but need not coincide with the objective consequence; the two may vary independently. But Laumann’s chosen referee is not just confused, his dogmatic social-psychological reductionism has made him invincibly obdurate.
Hickey’s paper says that an increase in the growth rate of the birth rate with a lag of four to eight years leads to an increase in the growth of compliance with criminal proscription of murder. Need it be said that this is not because it is the four-year-old to eight-year-old children who stop committing murders! If one insists on speculating about motives, one might say that the increased compliance is because the adult parents, whose domestic responsibilities motivate them to look to their own and their children’s futures, choose to be responsibly law-abiding parents. And it might also be because in this cyclical model the sixteen-year lag shows a negative relation to compliance, because teens are less socialized than adults. Criminologists know very well that national crime rates are greatly influenced by changes in the national demographic profile; teenagers are the foot soldiers in criminal gangs that engage in murderous turf wars.
Also the negative relation between changes in the per capita rate of compliance with criminal law prohibiting homicide and changes in the per capita rates of mass media exposure appears to corroborate the frequently expressed concern about the outcomes of favorable portrayals of violence in the mass communications media.
Hickey had made some computer runs with inputted data collected and released by the U.S. Department of Labor, Bureau of Labor Statistics that included women’s labor market participation, women’s employment and women’s unemployment. In those runs none of these variables were selected for output in any of the generated models. Hickey also inputted data collected and released by the U.S. Department of Health and Human Services, National Institute for Health, National Center for Health Statistics that inputted data included fertility rates and median age of first marriage. The discovery system did not select these for any of the outputted models, but instead selected crude birth rates. This critic’s claims about a “burgeoning literature” are too vague to be informative. Laumann’s chosen referee has not referenced any superior models containing variables representing the additional factors he demands. Hickey demands that Laumann’s chosen referee identify “concrete” and “specific” variables in empirical models that are more empirically adequate than those in his model.
Hickey’s so-called “post hoc” interpretations are the result of his recognition of relativized semantics determined by the linguistic context consisting of both the definitions described by the data’s source documents and by his empirically adequate equations estimated over fifty years of American history. The demand for “buttressing” independent evidence is gratuitous if not also cynical, because a demand for more evidence can be made of any paper at any time. It is indicative of the critic’s dogmatic mental state of denial in his attempt to invalidate the valid empirical model. Sending a writer after still more evidence is a well known sandbagging strategy. Similarly the demand for “substantively informed investigations” is code for social-psychological analyses. The evidence for Hickey’s model is its empirical adequacy, and his model is empirically adequate, even if this critic finds it “bizarre” in contrast to his dogmatic psychologistic prejudice. What this referee calls “post hoc”, Hickey calls “a posteriori”, which is to say “empirical”, which is to say “scientific”. Romantic sociologists like this referee are functionally illiterate in empirical science.
But Laumann’s chosen referee says he is “simply not convinced by anything” that Hickey reports. This obduracy is symptomatic of the mathematical illiteracy of a sociologist who is unconvinced because he is uncomprehending due to his incompetence. The problem is not with the model but with this dismayed referee’s inadequacy due to his obdurate psychologistic doctrinairism and incompetence in technique. Laumann’s referee criticisms have exposed American academic sociology to be an intellectual ghetto.
Laumann’s third criticism and Hickey’s rejoinders
Laumann: Laumann’s third chosen referee is the “internal referee”, which apparently therefore is internal to Laumann’s University of Chicago sociology department. Hickey believes the critic is probably Laumann himself, because his opinion is not independent. This critic says that he does not consider Hickey’s reply to the other criticisms to be sufficiently compelling to warrant reconsideration of the manuscript. He claims that Hickey “does not understand” the fundamental objection to the paper that was raised by both referees. He notes that both critics object to (1) the strategy of model building in the paper, (2) the interpretation of the variables (“indicators”) used in the models, and (3) the failure to provide a “convincing” sociological rationale for the specifications which the author settles upon – “specific mechanisms” in the words of one of the referees. The critic thus claims that Hickey’s reply does not satisfactorily rebut these three objections.
Hickey: The three criteria amount to politically correct ideas for the American Journal of Sociology under Laumann. This critic is just spouting more dogma in the “mechanisms” argot – the romantic social-psychological reduction seduction again. He is as obdurate as the other two, repeating the above criticisms and simply dismissing Hickey’s rejoinders. Claiming that Hickey “does not understand” is unmitigated arrogance. It is Laumann’s referee who obdurately refuses to understand.
In fact it may be said that Hickey did identify a “specific mechanism”, the integrative mechanism consisting of negative-feedback relations due to the interinstitutional cultural configuration of value orientations that pattern the propagation of social change through the system of types of institutional groups. Specifically it is the macrosocial stabilizing negative feedback relation due to increases in the high school completion rate, when there is sufficient economic growth. Notwithstanding that education is recognized as socialization, the macrosocial outcome is not obvious because it is intergenerational, and furthermore its effect is commingled with the interaction among all the other variables. But its operation was exhibited in the Type II simulation described in the paper. And it is a macrosociological interinstitutional relation and not a social-psychological relation that this doctrinaire critic will only accept as a “convincing sociological rationale”.
The first objection is just more Luddite ranting. The remaining two objections have been rejected in the history of macroeconomics, when economists unsuccessfully attempted to create a macroeconomics that is an extension of economic psychology, the maximizing rationality postulates of microeconomics. But economists recognized that macrolevel social analyses cannot successfully be reduced to microlevel individual analyses, because it incurs the logical fallacy of composition. Sociology is not an exception. In Parsons’ terms, macrosociology is “emergentist”, but contrary to Parsons it means that macrosociology cannot succeed as a reduction to an individualistic microlevel social psychology of motivational analysis. The social system has outcomes not possessed or exhibited by its component members.
But Laumann is a Parsonsian-era Harvard sociology graduate. He attended Harvard during the apogee years of Parsonsian influence with its requirement for motivational analyses that romantics like to call “mechanisms”. Any such deeply held vision comes to form part of the very identity of the believer, and any threat to the vision is experienced as a threat to the believer himself. Therefore Hickey is convinced that Laumann simply did not want a paper such as Hickey’s published anywhere much less in his American Journal of Sociology. He believes that the referee criticisms Laumann accepted reveal that he is an agenda-driven editor, that his choice of classical romantic referees was an ambush selection for a paper like Hickey’s, and that he got the kind of criticisms he wanted. Laumann’s reaction to Hickey’s paper is comparable to the “Last Sociologist” lament that Harvard’s Orlando Patterson published as an OP-ED in New York Times.
But outside the defensive ramparts of Laumann’s reactionary American Journal of Sociology things are changing. Recently as economists had seen in Keynes’ “paradox of thrift”, sociologists have recognized the logical fallacy of composition; just as houses need not have the rectangular shape of their component bricks, so too explanation of the macrosociety need not be in terms of the motivations of the society’s component individual members. For example in their “Quest for Institutional Recognition” in Social Forces (1998) sociologists Keith and Babchuk report that extension of individualistic microlevel social psychology, which they refer to as the “traditional individualistic modus operandi”, commits what they call the “individualistic fallacy.” These insights are not new. In the nineteenth century Durkheim, who viewed reductionism as threatening to sociology’s autonomy, went so far as to argue that whenever a social phenomenon is explained by a psychological phenomenon, the explanation is false.
Laumann’s choice of classical romantic sociologists for his referees brings to mind the report by Nobel-laureate economist Paul Samuelson, who wrote in Keynes General Theory: Reports of Three Decades that Keynes’ theory caught most economists under the age of thirty-five with the unexpected virulence of a disease first attacking and decimating an isolated tribe of South Sea islanders, while older economists were immune. This development was due to the pragmatism demanded by the Great Depression, when there was little patience with the doctrinairism of the conceited classical economists. Likewise Laumann’s classical sociologists are too ignorant of contemporary pragmatism and too inadequate in quantitative techniques – and too doctrinaire – to be inspired by the opportunities offered by new ideas, as were the young Keynesians.
Hickey expects that Laumann will impose
his social-psychological reductionism,
i.e., “specific mechanisms”,
on his sociology students for the remainder of Laumann’s
academic career. He
finds Laumann’s social-psychological reductionism suggestive of
movement advocated by the German historicist Wilhelm Dilthey and
his mid-nineteenth century sympathizers in their self-identified
Hickey believes that such classical romanticism has been so
retarding as effectively to be suicidal for the maturation of
academic sociology into an empirical science. Sociology at
University of Chicago was once pioneering. But Hickey counts
of referees among the “extinct volcanoes” that Harvard
University President Lawrence Summers would bypass for tenure.
American Sociological Review
The third academic sociological journal to which Hickey had sent his paper was American Sociological Review (ASR), the official journal of the American Sociological Association (ASA), which is published by the association and edited by a William H. Form at the University of Illinois, Urbana. Form acknowledged receipt of Hickey’s paper on 13 March 1981. On 10 April Hickey received a rejection letter signed by Form with two referee criticisms. Form edits his journal like a factory manager, who practices standardized production quality control. The resulting conformism results in publishing hackneyed outputs.
Form’s first criticism and Hickey’s rejoinders
Form: Form’s first chosen referee says that Hickey’s “metatheoretical considerations” do not motivate the actual analyses effectively, that little useful theory is involved, and that the particular analyses are similarly little motivated. He notes that none of them reflect the usually long traditions of research attempting to explain the variables involved. He called the paper “an empiricist venture” that is “utterly and ineffectively” related to the “empirical traditions” explaining the birth rate, homicide rate, etc.
Hickey: This referee’s reference to motivation is ambiguous. On the one hand whether or not a referee is adequately motivated is irrelevant to the validity of the model. If Hickey’s “metatheoretical considerations” do not motivate this referee to accept the paper, it is because the referee is innocent of contemporary pragmatism and inadequate to the mathematical techniques. On the other hand if the critic is demanding that Hickey describe motivations in his theory, the critic is apparently attempting to impose a social-psychological reductionist agenda on Hickey’s modeling work. The inadequacy of social psychological motivational analysis has been pointed out by many social scientists such as Robert Merton and Wesley Mitchell. The social-psychological reductionist agenda fails to recognize the existence of the distinctively macrosocial perspective that captures unintended and unforeseen – and thus unmotivated – consequences. Macrosociology is not just large-scale social psychology of motivations.
With respect to “useful theory” Hickey notes that utility is a sufficient condition for applied research, but is not a necessary condition in basic research. The mentality of Form’s chosen referee makes the referee unfit for basic research; he should have taken up occupational social work. Much basic research such as modern astronomical cosmology would be rejected as “useless” by this referee. In fact Hickey’s model is useful as basic research, because it is informative; it demonstrates empirically the strategic rôle both of rising per capita real income growth rates and of rising secondary-education completion rates for increasing macrosocial stabilizing consensus including rising voluntary conformity with criminal law proscribing homicide. The useful social policy implication of Hickey’s model is that increased public funding for universal public education together with progrowth Federal government macroeconomic policies increase macrosocial stability. Furthermore Hickey found his macrosociological model’s equation specifications useful for creating an Institutionalist macrosocio-econometric model of the American national economy for long-term economic-development policy analyses for the Division of Economic Analysis, Indiana Department of Commerce.
The appeal to tradition by Form’s chosen referee is truly appalling. It defies parody. It both reflects and explains academic sociology’s chronic decadence and protracted stagnation. Appeals to tradition are symptomatic of intellectual stasis, of lethargy due to indolent complacency, and of obstructing inertia of the comfortably familiar. Journals serving advancing empirical sciences seek to publish new and original research instead of tradition-bound hackwork. Sociology can never become a science, until it has become an empiricist venture. What is remarkable is not that science is an “empiricist venture”, but that sociologists like Form’s chosen referee need to be told that it is. This referee’s implication that sociology is not an empiricist venture explains why other sciences have demonstrated the progress that sociology has not, and why sociology exhibits its chronic legitimacy crisis.
Form: Form’s first chosen referee goes on to say that the empirical results reflect mainly trends in the variables, and that substantial secular trends are involved, so that some high correlations “naturally” result, which should not without much more thought be made the basis for causal inference.
Hickey: Had Form’s chosen referee actually attempted modeling, he would know that the task is not so easy as his dismissive rhetoric alleges. Hickey doubts that Form’s chosen referee even looked at the time-series data. Hickey had detrended the longitudinal data by transforming the per capita rates into growth ratios, to minimize collinearity among the models’ independent variables. These growth ratios have higher variances thus making empirically adequate modeling more difficult. This fact would be evident to an experienced modeler, but Hickey’s modeling is evidently beyond the competence of Form’s chosen referee.
Hickey’s models are causal models, not trend models. The critic’s phrase “more thought” is argot meaning speculation about motives, which is gratuitous. There is no need for any causal “inference” other than the satisfactory statistical inference employed. Hickey’s causal claims are based on the pragmatist thesis of ontological relativity with the semantics supplied by the context consisting of the descriptions in the data sources and the empirically adequate equations of the model. They do not represent fifty years of coincidence. But Form’s chosen referee is blissfully innocent of contemporary philosophy. He has attempted to force Hickey to submit the only kind of work the referee can understand.
Form’s second criticism and Hickey’s rejoinders
Form: Form’s second chosen referee wrote a shabby criticism containing many “X” overstrikes, and he appears to have given minimal time and superficial thought to Hickey’s paper. Form’s second chosen referee said that the paper is “ambitious”, and that its results are “questionable” and “meaningless”. The critic admits that while the estimated equations seem to make numerical sense in that they satisfy certain statistical criteria, they do not make “substantive sense”. He claims that demographers will be “amazed” with the finding that the decrease in the homicide rate causes an increase in the birth rate, that an increase in the birth rate twelve years earlier causes a decrease in the growth of the marriage rate, and that criminologists will be “surprised” to learn that an increase in the growth rate of the birth rate eight years earlier leads to a decrease in the growth rate of the homicide rate. He says that similar “puzzling findings” can be found in all the equations.
Hickey: Form’s second chosen referee’s term “meaningless” is uninformative except as an expression of disapproval. The semantics is exhibited by the descriptions of the data and by the equations of the model. Empirical findings are not wrong because they are unexpected – “amazing” or “surprising”. Saying that demographers will be “amazed” and that criminologists will be “surprised” is a cheap shot, which Hickey believes successfully panicked Form. In science amazing and surprising empirically adequate findings are characteristic of significant advances. Saying that the equations do not make “substantive sense” is uninformative and echoes Laumann’s referee who demanded “substantive reasoning”, a distinctively romantic verstehen thesis. It shows a failure to understand relativized semantics, the semantical thesis that is well known to linguists and contemporary philosophers of science.
Equation (1) determining changes in the crude birth rate says that changes in the crude birth rate are directly related to changes in compliance with criminal law prohibiting homicide. Increased compliance indicates increased macrosocial integration and social stability, whereas social disintegration and rising chaos discourage procreation.
This referee has merely glanced at the equations without considering the information exhibited by the simulation and shock analyses. Equation (2) says that changes in the per capita marriage rates are inversely related to changes in the crude birth rates eight to twelve years earlier. But in this cyclical model the marriage rate changes are therefore directly, i.e., positively related to birth rate changes twenty to twenty-four years earlier as exhibited in the type II simulation, because the model captures the life cycle in changes in the national demographic profile. The twentieth-century’s manifest demographic cycles are well known to demographers and even to some sociologists. They are more popularly known as the “post-war baby boom”, followed by the “birth dearth” or “generation X”, followed by the “echo from the baby boom” or generation Y, etc. By specifying the equations as they are, the sample data offer more degrees of freedom for statistical estimation.
Again this referee has merely glanced at the equations without considering the information exhibited by the simulation and shock analyses. Equation (3) determining changes in the per capita rates of compliance with criminal law proscribing homicide might “surprise” Form’s second chosen referee. But Norman Ryder, the demographer who published in American Sociological Review in 1965, would likely not be surprised that an increase in the growth rate of the birth rate with a lag of four to eight years leads to an increase in compliance with criminal law proscribing murder. It is certainly not the four-year-old to eight-year-old children who stop committing crimes, but as Aberle notes in America as a Mass Society the increased compliance is by their parents whose domestic responsibilities motivate them both to look to their own and their children’s futures and to be responsible law-abiding parents. Thus Ryder memorably wrote that nothing makes a young generation settle down more quickly than the younger generation coming up.
Furthermore in this cyclical model the sixteen-year delay in the change rate of the crude birth rates implies a negative relation to changes in the per capita rate of compliance with criminal law, because teens are less adequately socialized than are adults. Criminologists know very well that national crime rates are greatly influenced by changes in the national demographic profile, and that teenagers are less compliant, especially where they have supplied the foot soldiers for criminal gangs engaging in murderous turf wars. Also examination of research findings reported in the Federal Bureau of Investigation’s Uniform Crime Statistics reveals that the factors in Hickey’s equations would not “surprise” the informed criminologist, because several of Hickey’s factors are also in the FBI’s cross-sectional regressions.
To Form’s pontificating critic “ambitious” authors are heretics who must literally be excommunicated from sociology, i.e., denied an opportunity to communicate in the peer-reviewed literature. But to attack ambition is to defend mediocrity. This referee demands hack sociology although neither he nor Form would ever admit it. In empirical science there is no reason why a new theory should not be “ambitious” or why new findings should not “surprise” or even “amaze” cognizant professionals. And in progressive sciences (unlike hackneyed sociology) that is precisely what happens – and what gets published.
This critic’s demand that the model should make “substantive sense” is classic verstehen, hackwork in which familiarity – if not banality – is the criterion for scientific criticism, which in no small part has produced sociology’s chronic stagnation. For pragmatists and linguists the semantics is what one extracts from the linguistic context including the empirically adequate model, and not some presumption that one brings to the theory a priori. Thus Hickey echoes Lundberg who said that understanding is not a method of research, but rather is the end to which the methods aim. This referee’s criticism explains why sociology has been dismissed as merely “platitudes couched in jargon”. Hickey views this critic’s rejection of “surprising” and “puzzling” findings as professionally disreputable, because it is a disincentive for quantitative empirical research yielding new empirically “substantive” findings.
Form: Form’s second chosen referee claims that the main problem is that Hickey has not explicated a “theory” underlying the causal assertions embodied in these equations, and that the causal chains, if they exist at all, are very “loose and indirect”. He says that he does not deny that by computing suitable moving averages and growth rate indexes and by experimenting with lag structures, the author can find equations that seem to fit the historical record.
Hickey: What does the critic understand as “indirect”, much less “loose”? A criminal conviction for murder involves establishing a motive for the crime. But is the motive the direct cause of death? Is not the gun that the murderer used more direct than the murderer’s motive? But is the gun truly a direct cause? Is not the bullet discharged from the gun more direct? But is not the wound inflicted by the bullet the more direct cause of death? A physician could advance still more proximate causes thus revealing that “indirect” is an objectionable reductionist objection to the relations expressed in Hickey’s model.
Hickey’s empirically adequate model is revealing and sociologically significant. As shown in the simulations, the iterations of the model exhibit the propagation patterns of influence and their outcomes thereby revealing the net interinstitutional value reinforcement or conflict within the whole macrosocial system. There is more to understanding this dynamic model than just glancing at the regression estimated equations. And incidentally, since the equations in Hickey’s model are all monotonic, it is possible to express any variable as a function of any other in the model by simple algebraic substitution.
Form: Form’s chosen referee questions the “meaningfulness” of the equations, and adds that the root of this unclarity is Hickey’s “silly” idea of “theory”. He equates Hickey’s view of “theory” to saying that the empirical equations for the motion of a “spring-mass system” constitutes a theory that are distinct from the empirical equations for the “motion of a rocket leaving the earth”, because in both cases, the numerical equations are models of the phenomena in question deriving from the underlying theory of Newtonian mechanics. He says that a theory might help readers to accept as valid the causal specifications in the author’s model, which have very little face validity in the absence of such a theory.
Hickey: The statement by Form’s chosen referee that Hickey’s paper has no “theory” is odd given that functionalism has long been viewed by sociologists as a sociological theory, and the title of Hickey’s paper identifies the model as a post-classical quantitative-functionalist macrosociological theory. But Hickey denies that classical functionalist sociology is scientific theory, until the discourse is rendered testable and proposed for testing, as Hickey has done with his post-classical functionalist theory. Until it is empirically testable, traditional sociological functionalism continues to be just another descriptive grand narrative like practically all the sociological “theory” taught by sociology professors.
Like nineteenth-century positivists this critic has taken Newton’s mechanics as found in undergraduate textbooks as a paradigm for his concept of theory in research science. In the “Introduction” to his magisterial Patterns of Discovery: An Inquiry into the Conceptual Foundations of Science (1958), Yale University pragmatist philosopher of science Norwood Russell Hanson wrote that earlier philosophers of science had mistakenly regarded as paradigms of inquiry finished systems like Newton’s planetary mechanics instead of the unsettled, dynamic research sciences like contemporary microphysics. Hanson explains that such finished systems are no longer research sciences, although they were at one time. He states that distinctions applying to the finished systems, which he calls “catalogue science” as opposed to “research science”, ought to be suspect when transferred to research disciplines, and that such distinctions afford an artificial account of the activities in which Kepler, Galileo and Newton were actually engaged. He thus maintains that ideas such as theory, hypothesis, law, causality and principle if drawn from what he calls the finished “catalogue-sciences” found in undergraduate textbooks will ill prepare one for understanding research-science.
Hickey maintains that due to their almanac concepts of theory these sociology referees, not to mention the editors who selected them, are ill prepared to practice research science. Following Hanson’s functional pragmatism Hickey says that in research science “theory” is language at the cusp of scientific change in basic scientific research under the regulation of empirical criticism, the kind of change that constitutes productive work and enables scientific progress. But this referee’s “catalogue-science” view of theory has been an impediment to the realization of sociology as an empirical science.
Form’s second chosen referee says that both the empirical equations for the motion of a spring-mass system and the empirical equations for the motion of a rocket leaving the earth are numerical equations derived from the underlying Newtonian theory. This “catalogue-sciences” view of the meaning of “theory” suggests that of the sociologist Kenneth Land. In an article titled “Formal Theory” in Sociological Methodology (1971) Land offers a curious eclecticism that starts with the Hempel-Oppenheim positivist deductive-nomological concept of scientific explanation and ends up with Land’s version of the Haavelmo romantic structural-equation agenda for econometric modeling. In other words Land conceives theory as an organization of language as found in textbooks.
In fact his so-called “formal theory” is not formal. Its semantical interpretation is specific to his demographic-accounting agenda with its population stocks and “transition coefficients”. In his 1971 work Land explicitly defines scientific theory as a set of concepts and propositions asserting relationships among concepts instead of recognizing it as a transitory stage in achieving the aim of science. And he says that a distinguishing characteristic of a theory is that it cannot explain specific events without “prior transformation”, because theoretical propositions are general statements about causal relationships among the concepts in the theory, while there must be an observation record. Hickey comments that “theory” is not a special type of language, but rather is a special use of language; it is language that is proposed for testing in contrast to test design language that is presumed for testing. The pragmatics that defines theory is empirical testing.
Land then asks how to traverse the supposed “gap” between general theory and description of specific observations of empirical events. He thus assigned to empirical models the rôle that Hempel had assigned to empirical laws (before Hempel had read Quine’s “Two Dogmas of Empiricism”, and consequently in his “Theoretician’s Dilemma” in Minnesota Studies reconsidered positivism altogether). Land’s alleged “gap” is a pseudo problem. Functionally the defining feature of theory language is not its level of generality or its axiomatic organization. Rather it is the language that is proposed for testing, because its claims are judged to be relatively more hypothetical than the theory’s test-design language that is presumed for testing. The cognizant scientists thus agree that in the event of a falsifying test outcome the theory is the language that is in need of revision in contrast to the language describing the test design, although some scientists may revise their decision and practice counterinduction. And when the test outcome is falsification, then the theory is no longer a theory, but is merely rejected language. But if the test outcome is not falsification, then the theory is no longer so hypothetical as to function as theory language, because it has been tested, and the test outcome has made it an empirically warranted scientific law that can be used for scientific explanation and for test-design language for testing some other theory.
Land’s collaborator, Alex Michalos, has a similarly anachronistic view of theory. In his “Philosophy of Science” in The Culture of Science, Technology and Medicine (1980) Michalos references Hanson and calls Hanson’s thesis about research disciplines a “functional” as opposed to a “formal” view of scientific theory. This is a valid distinction, because Hanson’s pragmatic view is indeed functional for the practice of basic empirical research. But Michalos took the reactionary turn. In a “Prologue” co-authored with Land in Handbook of Social Indicators and Quality of Life (2012) edited by Land, Michalos and Sirgy, Michalos expresses his preference for the concept of theory as an axiomatic system, which is their concept of “formal theory”. Such “formal theory” is paradigmatic of what Hanson called “catalogue science”, because it is merely an organization of knowledge as an axiomatic system.
Like Land and Michalos, this critic has a similarly prosaic understanding of empirical science derived from the catalogue-science usage that is found in undergraduate textbooks describing the finished research findings of completed science, often completed many many years ago. The “theory” in the textbooks is like a museum taxidermy display of an animal’s stuffed carcass that cannot exhibit the repertory of animated behaviors of the living creature in its struggle for survival in the wild.
Therefore like Land and Michalos, Form’s second chosen referee is therefore ill-prepared for understanding research science and ill-prepared for practicing it, much less for contributing to its advancement. He and his ilk are destined to spend their careers in pursuit of a delusional Holy Grail for macrosociology based on their textbook stereotypes supplied by Newtonian mechanics. Contrary to this critic’s ridicule, Hickey’s view of scientific theory is not “silly”, and referring to it as such is an exhibition of egregious knownothingism. Hickey’s view of scientific theory is pragmatist since the pragmatics of theory language is empirical testing, which is what makes theory strategically functional for advancing empirical science in the practice of research science. Were sociologists to accept the contemporary pragmatist philosophy of science, they might do less wayward and ersatz philosophical dithering and more serious and productive modeling.
Hickey does not say in his paper that he used moving averages, although moving averages would serve his purpose adequately. The critic does not recognize that four-year moving averages over annual data would produce a smoothed annually incremented time series. Hickey used four-year incremented time periods for period averages. The use of four-year periods simplifies modeling, because there is then no imperative for complicated distributed-lag structures to relate a lagged-valued explanatory variable to the dependent variable of an equation, as would often be necessary with annually or shorter incremented time series. And the smoothing effect of the four-year periods also greatly mitigates outliers and removes noise from the data for which no model could account. While annual national-level data are already so aggregate that such noisy irregularities are usually negligible anyhow, the change ratios typically have higher variances than the measurements from which they are calculated.
In addition to simplification, there is also an incidental reason for the four-year period averages. One of the input variables to the discovery system is the political party affiliation of the U.S. President, who holds office for four-year terms. This was the only variable representing sociology’s conflict thesis that was inputted into the discovery system for developing the macrosociometric model, although it was not selected by the discovery system for any of the outputted models.
Finally saying that an underlying theory like Newton’s might help readers accept as valid the causal specifications in the Hickey’s mode is ironic, because Newton’s contemporaries, Leibniz and Huygens, had the same difficulty with Newton’s gravitational theory that this referee has with Hickey’s macrosociological theory. These contemporaries of Newton criticized Newton’s physics for admitting action at a distance; both Newton’s contemporaries were convinced that all physical change must occur through direct physical impact like colliding billiard balls, and Leibniz therefore rejected Newton’s concept of gravity as an “occult quantity”. Like this referee they too were unhappily “surprised” and “amazed”, and found the theory lacking what this referee calls “face validity”. In describing Newton’s theory as “unintelligible” Leibniz might have said that Newton fails to make “substantive sense”.
In his Concept of the Positron Hanson distinguishes three stages in the process of the evolution of a new concept of explanation; he calls them the black box, the gray-box, and the glass box. In the initial black-box stage, there is an algorithmic novelty, a new formalism, which is able to account for the phenomena. The equations in Hickey’s model are a black box for this referee. After some time scientists use this algorithmic novelty, but they then attempt to translate its results into the more familiar terms of the prevailing orthodoxy, in order to provide “understanding”. Romantic sociologists such as this referee could use the word “verstehen” here, even though he is too incompetent to use the formalism to understand the simulations. In the second stage, the gray-box stage, the new formalism makes superior predictions in comparison to the older alternative, but it is still viewed as offering no “understanding”. Nonetheless it is suspected as having some structure that is in common with the reality it predicts. This referee has not reached this stage. In the third and final stage, the glass-box stage, the success of the new theory will have so permeated the operation and techniques of the body of the science that its structure will also appear as the proper pattern of scientific inquiry. For example in the nineteenth century Helmholtz wrote that Newton’s theory has become the paradigm of explanation in physics; it had become what Hanson called a glass box. With Einstein’s general relativity theory, gravity has been reconceptualized again, and Newton’s glass box was as it were broken except for those reactionary physicists who continued to find Einstein’s theory a black box. Today Keynes’ macroeconomics has the status of a glass box for economists. But Hickey’s post-classical functionalist macrosociometric model is still a black box for the reactionary referees selected by Form and the other editors.
Form: Form’s second chosen referee accuses Hickey of distorting Land’s position, and claims that Land did not argue that it is imperative to take the equation specifications of sociological models from existing colloquially expressed theory, but rather just stated realistically that this is the typical level of most current sociological theory. With suspiciously miraculous clairvoyance the critic reports that Land would be the first to admit that a theory can be mathematically stated. For a general theory he again references Newtonian theory, as a paradigm that he says would be much better, if such could be constructed for sociology.
Hickey: This “distortion” rhetoric suggests that the critic really knows how to yo-yo an editor. In fact Hickey’s issue is not about Land’s views on mathematically vs. colloquially expressed language, although the critic is correct in saying that colloquially expressed theory is how romantic sociology with its motivational descriptions is typically expressed. Hickey’s issue is about the source and justification of the equation specification for models, i.e., of the selection of explanatory variables for any model, which is the central issue for this referee, and its semantical interpretation. Hickey affirms that extraction by data mining is a valid source for equation specification, that empirical adequacy is adequate justification, and that its interpretation is by relativized semantics.
In his “Social Indicators” article in the Annual Review of Sociology (1983) Land disapprovingly states that analysis of demographic time series have employed ad hoc combinations of econometric and time-series techniques that ignore underlying population dynamics of the social-indicator time series. Hickey finds this description suggestive of Land’s attempts at macrosociological modeling in 1975, which depends on autoregressive variables. Autoregressive variables capture the effects of missing explanatory variables, and thus are a way to make bad equations look statistically acceptable. The equations in Hickey’s macrosociological model contain no autoregressive variables.
Contrary both to Land and Michalos and to Form’s second chosen
referee empirical testing is the
pragmatics of theory language, and is the defining
characteristic of theory in research science; thus the theory
is the model and the model is the theory.
But in his 1971 paper
Land distinguishes theory and statistical model, and gave no
hint that he would accept data mining as the source for the
equation specifications, accept empirical adequacy as their
justification, or accept relativized semantical interpretation. Meanwhile longitudinal
modeling in sociology stalls.