Stephen Chilton


A previous version of this article was presented at the panel, "Strategic Thought, Political Action, and Public Policy", at the annual meeting of the Midwest Political Science Association, Chicago, April 13, 1989. I am indebted to Professors Craig Grau and Lois Sayrs for encouragement and/or intelligent commentary. I remain, of course, solely responsible for any lapses of judgment.


This work lays out the logical framework of one form of intellectual discourse, here termed a "hermetic project", in which researchers are induced to study a series of successively falsified theories whose common structure may not allow for accurate representation of the truth. Such frameworks I term "hermetic" because they are logically consistent and self-reproducing but not capable of self-transcendence. The term "project" connotes the organized, goal-seeking nature of the traditions characterized by hermetic thought. The image of a goal-seeking organization that is at the same time trapped within the limits of its reasoning system -- that image is exactly what this work seeks to convey and the reason for our concern that we transcend it.

The intellectual limitations of such projects have negative effects in many different fields. In addition to describing the nature of hermetic projects, the paper outlines the operation and intellectual structure of three such projects: the "liberal project", the "conservative project", and the "fault tree safety studies project".


"You couldn't have it if you did want it," the Queen said. "The rule is jam tomorrow and jam yesterday - but never jam today."
"It must come sometimes to jam today," Alice objected.
"No it can't," said the Queen. "It's jam every other day: today isn't any other day, you know."
- Lewis Carroll Through the Looking Glass (19xx: xx)
"We seem to be at cross-purposes," he said. "I send you down to discover why Fennan shot himself. You come back and say he didn't. We're not policemen, Smiley."
"No. I sometimes wonder what we are."
- John LeCarré Call for the Dead (Heinemann: NY, 1979), p.192


Hermetic projects are unfortunate conjunctions of methodological, sociological, and psychological structures that, while logically consistent and self-reproducing, are incapable of self-transcendence: thus their characterization as "hermetic". "Project" refers to the organized, goal-seeking nature of the traditions characterized as hermetic. Taken together, these two terms -- "hermetic" and "project" -- convey the image of a goal- seeking organization trapped within and frustrated by the limits of its reasoning system.

Since hermetic projects appear in many different fields, their limitations have pervasive negative effects. As the examples presented below will show, hermetic reasoning both structures and thereby limits much of our current scientific, technical, social, and political life. In other words, the reasoning structure creates now, and may yet create, much misery.

The concept of a hermetic project is obviously related to that of paradigm (Kuhn 1970). Briefly, one could say that hermetic projects constitute cultures by virtue of their participants relating to one another in terms of the paradigm. As will be seen later, this view supports Kuhn's. The present paper, however, is more concerned with identifying the abstract reasoning/structural basis of paradigm in order to apply it to other fields.

My general goal is to devise ways for researchers to recognize and move beyond hermetic projects; this paper represents only the first step toward this goal, and an uncertain one at that: a preliminary attempt to outline their structure and operation. If readers find that my logical cart wobbles a bit, I encourage them to point out which wheel has fallen off; I hope they will also be able to tell me where to find a better one.


Before dealing with the four cases to be examined, it will be helpful to survey the landscape we are about to enter. First, the heart of each hermetic project is a rational, scientific core of reasoning. This core is fundamentally the logic of the scientific method. Each hermetic project is based on two different categories of mull hypotheses. The first category is one we are familiar with: the falsifiable statement to the effect, "Such-and-such a specific theory hold in this case [or, in other contexts, such-and-such a practice accomplishes our goals]." We will label such null hypotheses of the first category "h0", using a lower-case H.

Because these null hypotheses are falsifiable, any research tradition will have postulated and falsified a string of them, probably of increasing complexity and presumably becoming closer to the truth in some sense. "Problem- solving", "normal" science consists of the applications of the current null hypothesis h0 to a new situation. Logically, if the null hypothesis is generally correct, then it must apply to this case in particular. Thus engineers build bridges using standard engineering practices (which include mathematical computations and analysis of stress and strain according to standard theories, as well as the practice of what bridges ought to look like), fully expecting the old practices to work in the new case. It is this sort of extension which Kuhn (1940:174-5) calls the "problem-solving" meaning of his term "paradigm."

A problem arises when h0 is falsified (e.g., when the bridge collapses). This violation of the null hypothesis (clearly our practice is not accomplishing our goals of having the bridge stay up) implies that we must formulate a new null hypothesis. This is the ambit of the second category of null hypothesis: statements about what form our alternative hypotheses can take. For example, implicit in Newtonian physics was a concept of absolute space, a concept recognized by Newton as problematic but assumed thenceforth by his followers to be true. We will label null hypotheses of this second category "H0", using a capital H.

A central issue of the philosophy of science is under what circumstances the disproof of h0 implies a disproof of H0. Disproof of the former may call the latter into question, but it does not disprove it: new h0's spring up from the ground of H0, as Antaeus did when Hercules threw him to his mother, Earth. Even the lack of any reasonable h0 does not disprove H0: one can attribute the failure of h0 to experimental error, or one can hold that a new, as-yet- unknown h0 will appear. Such arguments are not totally convincing (taken to the extreme, they prevent theories from ever being falsified), but on the other hand this logical weakness of the argument does not prove it to be untrue. As Kuhn (1970:10-34) points out, workable theories are created by extension from an initially simple formulation.

The above points are the aspects of logic with which we will be concerned in this paper. There are, in addition, points to be made about the psychological and sociological aspects of hermetic projects. The individuals involved in the hermetic projects are psychologically invested in them. Long training and experience in a project makes it psychologically familiar and comfortable. Additionally, the participants in the project constitute an exclusive community which, by its ability to control social support, provides yet more incentives to maintain the project. Taken together, these factors may protect H0 absolutely. In the absence of any clear standard of proof or disproof, it will continue to order intellectual discourse far beyond whatever criteria philosophers of science decide it should meet.

In summary, we consider the following aspects of hermetic projects: (1) the null hypothesis h0; (2) the application of the null hypothesis to problems (the "problem-solving" meaning of paradigm); (3) the application of more general considerations (the "sociological" or paradigmatic hypothesis H0) to the revision of the null hypothesis; (4) the rejection of apparent anomalies as experimental error or as "resolvable in due course"; (5) psychological investment of the participants in the project; and (6) sociological support by the community of participants for the project.


Our first example of a hermetic project is taken from "fault tree" analysis of nuclear reactors to determine their safety.

3.1 Fault Tree Analysis

A fault tree analysis is in theory no more than an enumeration of all possible ways in which the reactor might fail and a computation of the likelihood of these various possibilities. A simple, hypothetical example is shown in Figure 1 below.


Figure 1 about here


The analysis is called a fault tree analysis because of the branching of possibilities as we go farther in time into the possible failure: first we consider the possible states of the primary cooling system, second the possible states of the back-up cooling system. The branches, down to the last twig, each represent some particular possibility for how parts of the reactor will behave. In Figure 1, it can be seen that reactor failure only occurs after failure of the primary cooling system and failure of the secondary cooling system. Each of these possibilities, taken alone, has a certain probability of occurring: Figure 1 assumes there is a failure probability of 1/1000 (.001) for the primary and 2/1000 (.002) for the back-up cooling system. Presumably these figures would be obtained in real analysis from previous reliability studies of these systems. The probability that the reactor will fail is therefore .001 * .002, or .000002 (i.e., 2/1,000,000), assuming that the status of the primary cooling system is not related to the status of the back-up: they operate or fail independently of one another. In Figure 1, there is only one route to failure, and it has a probability of .000002 of occurring. If there had been more routes, we would have summed the probabilities of all failure routes in order to get the overall probability of failure.

3.2 Two Levels of Analysis and Logic

The reactor safety project is based on two related hypotheses. The null hypothesis h0 is, "This fault tree model of the reactor is correct." This hypothesis allows post hoc analysis of reactor behavior; any reactor incident can be analyzed against the model to determine if the model's assumptions are met. Generally, falsification of h0 is straightforward: while some reactor incidents will follow the paths and assumptions shown in the fault tree model, others will not. For example, the fire among the control cables at the Browns Ferry 1 reactor disabled several systems simultaneously (Ford 1982:82 ff.). Such circumstances are termed "common-mode failures," and they violate the fault-tree model's assumption of independence. The falsity of h0 -- that is, of the specific fault-tree model -- is accordingly quite clear. One can therefore change either the model (to reflect the possibility of that specific failure's recurrence) or the reactor (to prevent the recurrence). The first option is unacceptable, of course, because the aim is to prevent failures, not to model them. The second option is the one taken in general: in the specific case of Browns Ferry 1, the Nuclear Regulatory Commission (NRC) required the subsequent use of cables with fire-resistant insulation.

The second hypothesis -- the metatheoretical or paradigmatic H0 -- is an a priori statement about the reactor. The hypothesis at the level of theory, h0, is a contingent one: "Let's pretend this is the way the reactor works and see if in fact it does." The hypothesis at the level of metatheory, H0, is not contingent at all: "This model, right now, is correct" or, more broadly, "It is possible to estimate reactor safety in advance using fault-tree analysis." As we saw above, it is easy to challenge h0 -- indeed, it invites challenge; it is constructed to be challenged. But H0 has a different logical status. h0 is designed to be tested after the facts are in: did the reactor behave as we said it would? H0, however, cannot be challenged in that fashion, because it has to do with an event not yet observed: the ultimate adequacy of this h0 (or the eventual adequacy of some h0). Our scientific concern is with h0 alone: we are happy to challenge, defeat and rebuild h0 as often as necessary. Our social concern, however, is with H0: we want to know right now, preferably before the reactor is even built, that it will be safe.

Failure to recognize the distinction between h0 and H0 is a source of difficulty in arguments over reactor safety. The socially significant argument should concern H0. It should pit against each other a confidence in the ability of the governmental-technical-economic system to carry out the construction of a reliable complex system, on the one hand, and a doubt about the ability, on the other. In practice, however, the debate is not between H0 and H1 but instead between h0 and its negation, which is only of technical concern. Safety analyses are carried out, each one constituting a new h0, and are attacked either theoretically (they are shown to ignore certain factors) or by demonstration (a reactor incident occurs which violates their assumptions). In either case, a new h0 is ready to replace the old one. It takes time and/or effort to disprove any given h0, and it is logically impossible to prove that all such h0's are wrong. (After all, some fault tree has to be right.) So at a scientific level of argument -- at a level of discourse which permits only empirically or analytically provable statements -- H0 cannot be challenged.

The arguments over reactor safety are maintained at the level of h0 by the lack of public attention except when dramatic, concrete events like Three Mile Island have occurred. Efforts to broaden consideration of the specific problem in the context of a general root problem (i.e., to pass from consideration of h0 to consideration of H0) are branded "radical," unconsciously following once more the linguistic path from the Latin word "radix," meaning root, to its modern derivative. Because the problem is narrowed to the specific incident (that is, to the failure of the analyst to predict or prevent the specific incident), the solutions remain specific. The mathematical model of the reactor is compared with the specific events occurring in the incident, and the inevitable discrepancies are compared. (I say "inevitable" because if there were no discrepancies, then the incident would be very unlikely to have occurred, given the minuscule probabilities shown by safety analyses.) The problem is thus transformed from "How if at all can reactors be modeled?" to "How can the model be changed to show the possibility of this event's occurrence?" Solutions to this latter problem come easy: Did operators turn off the wrong switch? -- then add an "operator-turns-the-wrong-switch" branch to the event tree. Does this result in an unexpectedly high probability of an incident as happened? -- well, by jingo it does! So how can we keep that incident from occurring? -- well, just eliminate the operator's discretion, or make the switches automatic, or instruct the operators better, or, or.... And it's true -- Three Mile Island will never ever happen again. Which is to say only that the exact set of events and circumstances which caused the incident will not occur again. Like generals, safety analysts are always fighting the last war.

So the safety analyst's mode of thought goes like this: (1) "My mathematical model shows me what happens in the reactor." (2) "If an incident occurs which I didn't foresee, then the model must be wrong." (3) "I can correct the model by comparing it to the incident and making the indicated adjustments."

But from a citizen's point of view the problem is not how to make a better model but rather how to make a safe reactor. A revised model does not ensure that the reactor is safe.

3.3 Psychological and Sociological Patterns

Once one recognizes the distinction between h0 and H0 and also recognizes the alternative to H0 (i.e., H1), then one must also recognize that there is no reason why we must take H0 as our null hypothesis. To use a legal analogy, we assume H0 to be innocent until proven guilty (i.e., we evaluate it in terms of criminal law) instead of evaluating H0 against H1 on the basis of the preponderance of evidence (i.e., evaluating it in terms of civil law).

How has this state of affairs arisen, and what supports it? From a purely scientific point of view either H0 or H1 can be used as a starting point, and thus it is a matter of scientific indifference which of the two is used. The fact that H0 continues to act as our null hypothesis therefore must be explained apart from issues of scientific logic. To do so we examine its origins and the psychological and sociological meaning of its continued existence.

Science and its children, technology and engineering, have had for over two hundred years a relatively unbroken string of triumphs. Both in terms of our abstract understanding of nature and our position in it and also in terms of our concrete capacity to make the world over to serve our wishes, science has affected us enormously and, it seems to us, for the most part, positively. We did put a man on the moon -- several, in fact. There is no gainsaying that science knows much and that technology and engineering have achieved much. To say we are capable of achieving this or that technological goal is thus no more than to say we are capable of repeating the feats of the past. Nuclear reactors are not moon flights, to be sure, but the analogy is not at all unreasonable. Certainly the orientation represented by H0 has clear historical justification.

From a psychological perspective, people who adopt H0 achieve a number of goals. First, they identify with what seems to be a proven winner. They look for technological successes, and they are enlarged by them. Second, to the people who actually engage in reactor design, their professional life has been built around the possibility of building a reactor. To adopt H1 rather than H0 would in effect be saying at a psychological level that their life had been without meaning: that they had devoted their life to pursuing a will-o-the-wisp. Meaning is constituted in intentional action, and to assume H1 would be to assume that one's life of designing and building reactors has been meaningless.

It is equally evident that H0 is grounded in a sociological network, or, conversely, that H0 is the organizing principle for a sociological network. As we look at the scientific-technological enterprise founded upon H0, we can see how interconnected it is: wherever we start, we can traverse the entire network of influences. (1) Start with an engineer's career. In the nuclear business, it is predicated upon the assumption that nuclear reactors can be built (i.e., upon H0): nuclear firms have no use for critics. Critics of specific reactor designs (i.e., of h0) may be accepted as productive of safer designs, but they must cut their criticisms off at some point. Critics of H0 itself cannot be accepted: to reject H0 is essentially to reject the entire enterprise upon which the firm is founded. (2) Start with an academic's career. In the area of reactor safety, journals and funding agencies are controlled by people with a belief in H0 -- a belief which makes possible the very enterprise in which they are engaged. Why fund reactor safety studies if one knows a safe reactor can't be built? Why study reactor safety at all -- Why write about it? Why teach it?

H0 is a far-reaching organizing principle. Its adoption gives meaning to what people do (in fact, it is the meaning of what people do), and this individual adoption has, over time, been elaborated into a complex sociological network. To reject H0 is to undercut at a stroke the cultural meaning on which this enormous, powerful edifice is raised.

3.4 Summary

In summary, dealing with a never-ending succession of h0's prevents us from dealing with H0, their source. While any specific h0 is subject to scientific testing, H0 is not: H0 represents a paradigmatic, a priori view of the world. In disputes between H0 and H1, our criterion of proof or disproof is not scientific; rather, since H0 and H1 are ways of making a priori sense of the world, they must be decided in a process of discourse. H0 is grounded in personal meaning as well as meaning derived from a sociological network. This claim is not to negate in advance all arguments by supporters of H0 but to enable them to recognize and transcend, in their self-examination and in their communication of their knowledge to others, arguments deriving solely from these sources. If all parties can recognize in principle that they are situated within a certain framework of meaning, they can in practice evaluate the significance of their experience in terms applicable to all.

[December 16, 2000:  I want to add somewhere that my initial idea about hermetic projects came from reading, probably back in the early 1980s, Donald Knuth's The Art of Computer Programming, in which he discussed the problem of when to rewrite an operating system instead of merely continuing to patch it.  He shows that empirically, at least, the number of known bugs in the system decreases from the time of its release but then after some period of time (which I have forgotten now -- probably around 3 years) begins to rise again.  In other words, the patches introduced to fix old bugs begin to interact with each other and the system, so that each new patch for one bug creates on the average more than one other bug.  It is at that point that the system designers need to consider bringing out a new system entirely, even though that new system will initially have more bugs than the system it replaces.  The connection to my reactor safety studies example, above, should be obvious.]

[December 16, 2000:  I also need to consider referencing James H. Fetzer (1998?).  Computer Systems:  The Uncertainty of their Reliability.  Bridges 5(3/4).  Pp.197-217.  I think he makes a similar point, or at least one that would support my argument, but I need to talk with him about it.]


A similar problem occurs in science as a whole, as Thomas Kuhn (1970) pointed out in his distinction between "normal science" and "revolutionary science." Normal science, like reactor safety studies, is a hermetic project. At the risk of some boredom due to the similarity of this example to that of reactor safety studies, let us examine the logic of science.

To begin with, we distinguish between "current theory" and "paradigm." The former is our h0 -- our null hypothesis at the level of formal logic, of scientific proof. It says that this particular theory is correct in all its details, and it is tested on the a posteriori basis of its successful prediction of certain experimental results. Falsification of h0 results in its rejection, of course, and its replacement by a new h0 which could explain the anomalous result. What are the characteristics of the new h0? It is Kuhn's point that alternative hypotheses are, during periods of "normal science," all consistent with a more general framework. Kuhn calls this framework a paradigm; I call it H0. As in the case of reactor safety studies, H0 is a priori statement.  In advance of our knowing what theories are really true, we say that "the really true theory must have (such and such) characteristics." In the classic example of Ptolemaic astronomy, H0 said that the motion of celestial bodies was based on certain cycles and epicycles, and the project of normal Ptolemaic astronomy therefore became to determine exactly which set of cycles and epicycles was the true set.

At the level of normal science, such a logical structure is indeed hermetic: impervious to falsification and incapable of self-transcendence. The rejection after painful analysis and observation of one set of cycles merely results in the proposal of another set. As long as one stays at the level of falsifiable hypotheses, one neither tests nor rejects the paradigmatic cornucopia of h0's: H0 itself. In a classical statement, Sir Karl Popper decried Kuhn's conception of science:

In my view the "normal" scientist, as Kuhn describes him, is a person one ought to be sorry for.... The "normal" scientist ... has been badly taught. He has been taught in a dogmatic spirit: he is a victim of indoctrination. He has learned a technique which can be applied without asking for the reason why....
But such a statement is tenable only in the abstract.  It fails to consider that humans are guided by a priori judgments about what routes are likely to be profitable. Popper may not like to believe that we are guided by such beliefs, but it certainly appears to be what science (normal science, anyway) is like. True, our a priori are not susceptible to scientific testing -- they are not scientific at all in the narrow sense -- but this does not seem cause for despair. These a priori hypotheses are useful in our selection of reasonable new h0's once old h0's are set aside. Just as miners do not strike off at random in their mines but rather follow a seam of ore until it appears to be played out, so scientists do not normally utilize non-H0 alternative hypotheses until the generation of hypotheses within H0 appears to be profitless. Not to employ an H0 would essentially mean to abandon all hope of doing science.

It could be argued that we are not mining ore but rather fool's gold when we follow a false H0 -- that all the presence of H0 does is lead us astray for a longer time and in a more concerted way than otherwise. But while a paradigm shift can involve a massive reconceptualization of facts (including the revelation that many facts were only artifacts and that many puzzling results were not problematic at all), most facts remain constant. For example, relativity theory reinterpreted but did not invalidate most experimental results based on Newtonian mechanics: the scientific effort since Newton was not wasted but rather placed in a broader context. Using paradigms gives research a direction, focus and meaning which is never entirely negated by the adoption of a new paradigm.

These, then, are the logical considerations surrounding the use of h0 and H0 in science. It remains to examine their psychological concomitants. Since these are similar to those discussed in the previous section, the discussion here will be brief. Psychologically, scientists are reluctant to abandon theories to which they have devoted years of their lives and which form the deep roots of their understanding of the world. As Kuhn (1970:76) puts it: "As in manufacture, so in sciences -- retooling is an extravagance to be reserved for the occasion that demands it." To lose one's belief in a given paradigm is to experience a sense of meaningless confusion, and possibly despair. Kuhn (1970:84) quotes Wolfgang Pauli's words during a crisis in quantum mechanics: "At the moment physics is again terribly confused. In any case, it is too difficult for me, and I wish I had been a movie comedian or something of that sort and had never heard of physics." Or, quoting Einstein, "It was as if the ground had been pulled out from under me, with no firm foundation to be seen anywhere, upon which one could have built." Unless a person is very firmly grounded in an understanding of science that recognizes the existence of paradigm shifts, the feelings of confusion or meaninglessness can easily be intolerable. They are never pleasant, even if they are understood. It would therefore be tantamount (on an intellectual plane) to committing suicide for one to abandon one's paradigm. Like jumping out of a burning building from the fourth floor, one does so only when the flames are upon one. Theories are also surrounded by and embedded within a sociological infrastructure that constitutes a social control mechanism. Without claiming that each one of these institutions is abjectly subservient to a ruling paradigm, one can still point to the following aspects of a paradigm's institutional life which, taken together, serve to make research within the paradigm easy and rewarding and to make other research difficult and unrewarding:

None of these barriers to a free professional life is impenetrable. Rather, their effect is to make work outside the paradigm enormously more difficult: instead of walking through a series of open doors, the researcher outside the paradigm faces obstacles at every turn. These factors (in addition to the psychological factors discussed above) lend a non-logical weight to the logical considerations mentioned above. In consequence, scientific research paradigms are not abandoned by current scientists but are, rather, supplanted by new generations' adoptions of new paradigms. Within the intellectual life of a given scientist, as in the reactor safety studies, an empirical challenge to the existing theory is met by minor revisions within the original paradigm. Fundamental (radical) questions are not asked. What at one level (the level of normal science) is seen as a process of continual challenge and revision, at another level (the paradigmatic level) is seen as a hermetic process of self-perpetuation.


Let us turn now from science to the world of social/ethical thought, looking at two hermetic projects: what will be termed "the liberal project" and "the conservative project." The liberal project here refers to attempts to correct social inequalities through a configuration of government- administered or -supported compensatory programs.

Viewed at one level, the configuration of programs undergoes a continual process of improvement in response to inadequacies found in it. In an ideal-typical sense, the following sequence occurs:

(1) It is determined that certain people are failing to benefit (or are benefitting less than others) from the existing social arrangements, recognizing both general socio-economic institutions and the existing configuration of compensatory programs. In liberal parlance, they have "fallen through the cracks." The existence of such people constitutes a challenge to the liberal project's h0: that the current society (again, recognizing that it includes both general socio-economic institutions and the existing configuration of compensatory programs) allows all people to achieve a certain minimum standard of living. Our hypothesis h0 that people are not falling into poverty (or, to phrase it more in the liberal spirit, that people are being lifted out of poverty) is tested a posteriori by an examination of whether people in poverty still exist.

(2) Brought into play at this point is the liberal project's H0: the a priori assumption that society should guarantee all people a certain standard of living.

(3) The liberal analyst applies this perspective by examining the persons still in poverty with the view to discovering what social force prevented them from utilizing the resources available in the society. Some people, of course, simply choose not to use compensatory programs, denouncing them as "socialistic" and/or pointing with pride to how their "family never went on the dole during the Depression." These people do not constitute a challenge to h0. They have so clearly chosen poverty over the liberal alternative that attempts to get them to use the programs would constitute interference with their right to self-determination. Since (to the liberal) there is nothing wrong with taking advantage of compensatory programs (after all, they're there because you have a right to a certain standard of living), objections to "the dole" only appear misguided and old-fashioned: admirable in their way, of course, but out of place in the modern world. Under these circumstances, no change in h0 is required. However, many individuals do directly present a challenge to the liberal h0. Poor people are refused various benefits and lack the money to take legal action. Many elderly live in rural areas and lack transportation to get to centrally located senior citizen centers. The examples can be multiplied endlessly. In each case, however, it is clear that the people involved do not have access to the facilities which h0 presumes they do. Thus h0 is rejected.

(4) According to H0, any HEW h0 must have the form of a new configuration of compensatory programs, and the search for this new h0 organizes the political arena within the liberal camp. The issues become what new configuration is to be instituted -- what new deprived group (e.g., "rural elderly") or new need (e.g., "legal services") has been revealed in the rejection of h0. What new program, or revision of an old program, or restructuring of the configuration of programs, is needed to solve this problem? The outcome of such a political debate (which of course eventually meets and is modified by its confrontation with the conservative position that no programs are needed at all) is a new h0, a new configuration of compensatory programs in interaction with the socio-political system.

It is at the closure of this cycle, just prior to beginning the identification of new problems as reasons for yet other programs, that the liberal project itself can come under challenge. It is not under challenge during the other steps of the cycle, because debates over the form of the specific projects still contemplate social programs as solutions for problems. No, the time for challenge comes when new configuration of compensatory programs have produced new problems.

How is it that the problems produced by the old cycle do not cause the liberal project itself to be abandoned? As in the previous discussions, we turn to logical, psychological and sociological factors for our answer. Logically, the problem once again is that H0 is an a priori hypothesis, not a scientific hypothesis. Scientific hypotheses can be challenged by a posteriori counter-examples, but H0 speaks of direction in which in advance of any evidence it is thought fruitful to go. One can choose to believe H0 or not, but the commitment is made on other than scientific grounds. William Ryan (1976:251) compares it to a spotlight: "We look where the spotlight shines: one of the first tasks in breaking the ideological spell ... is to shift the beam of the spotlight."

Psychologically, the liberal project is founded on a recognition that social institutions are mostly arbitrary (that is, not moral merely because they are institutions, as a "law and order" perspective would have us believe), a recognition of the agency of past social institutions in creating systematically deprived classes of people, and a recognition of the self-serving potential of a blame orientation.

Sociologically, the programs created by the liberal project take on a life of their own, socializing their employees into the perspective of the liberal project and providing power bases for political activists who share this perspective. All of these forces provide incentives to fix old problems with new programs rather than to ask the root (radical) questions about possible wider, common causes of these problems or about their amenability to "solution" at all.


I turn now to my fourth and final example, which is just the obverse of the previous one: the politics of blame, or, lumping together in a Procrustean fashion a variety of political philosophies, the conservative project. As in the liberal project, the existence of poor people poses a problem for the conservative project. In both projects there is a basic discomfort (as there should be, of course!) about the existence of poverty. However, the two projects interpret the discomfort in different ways, each according to the null hypothesis and especially the underlying paradigmatic hypothesis H0. For the conservative project, h0 is the hypothesis "This society provides equal opportunity for reasonable, prudent people to succeed," and the underlying hypothesis H0 is "Equal opportunity for reasonable, prudent people is all one is morally obligated to provide." The liberal project accepts these hypotheses, but adds the condition to h0 that everyone can have some minimum standard of living and to H0 that we need only provide such minimum standard of living (in addition to the equal opportunity).

Poverty (or any other inequality), therefore, challenges the conservative project's h0 by calling into question the equality of opportunity in the society. The conservative project meets this test by asking whether the poor are behaving reasonably and prudently. It asks, in other words, "In what way are the people having problems to blame for their problems?" anticipating and indeed calling forth an answer of the form "These people are failing to do such-and- such a thing (e.g., studying and working hard; deferring gratification; obeying the law) which reasonable, prudent people do." They could have succeeded, so the argument goes, if they had just done what was expected.

Such an answer is immensely powerful. First, it is difficult to press one's case when one is at fault for one's troubles. Such an answer plays upon every man's recognition of his or her own lapses, enfeebling the protest within ourselves even before it can be made public. Second, the assistance which is given in such circumstances is not given from a sense of justice or of love; rather, it is given as an opportunity to play Lady Bountiful (and the peasants should be humbly grateful) or in anticipatory fear of a peasants' uprising (and so the relationship is one of resentment of the extortion). These issues are a major source of welfare system problems. Third, the argument negates virtually all claims. Since none of us are perfect, it is extremely difficult to convince someone that our problems are not due to our lapses. It is this powerful ability to set aside claims which I believe is at the root of the political alliance between the two forms of conservatism: the one form wishes to see no change from tradition, while the other has a philosophical stance which almost never compels change.

Despite this, there are cases in which h0 can be rejected, cases in which evidence of discrimination, of systematic inequities is very clear and widely documented. In these cases the conservative project does reject h0: it concludes that equality of opportunity has been violated, it supports the indicated change, and it thus re-establishes a new h0: "This society provides equal opportunity for reasonable and prudent people to succeed."

Left untouched is H0: "Equal opportunity for reasonable, prudent people is all society is morally obligated to provide." Thus, as with the previous examples, the problems represented by certain circumstances (previously: a reactor incident, an inexplicable scientific fact, the existence of people in difficulties despite social programs; here: people in difficulties despite equality of opportunity) are overcome in a manner which leaves untouched the underlying assumptions. The existence of desperately poor people is simply not a matter of concern in the conservative project except insofar as it represents the results of inequality of opportunity.

How is it that the continuing presence of the poor does not cause the conservative project to be abandoned? To lay this out fully, we review its logical, psychological and sociological underpinnings. In a narrow logical sense, the conservative project is self-consistent, which is to say it is hermetic. Having excluded all issues except equality of opportunity, it is not logically bound to consider other issues like inequality of outcome. In a broader logical sense, that is, attempting to look at the sense of H0 directly, we can see that there is a value to encouraging reasonable, prudent thought, and one would like to be able to plan and create institutions which can assume this of people. Such institutions will be more productive, more effective, more efficient than institutions which cannot make such assumptions. And people do have some choice about being "reasonable and prudent," however this phrase may be defined. So one cannot simply dismiss the conservative project as barren of truth.

Let us turn to conjectures about the conservative project's psychological and sociological bases. Psychologically, it is pleasant to be able to attribute one's own success to one's innate moral worthiness -- to have evidence by one's success that one is among the elect. These feelings have been studied and discussed extensively by McClelland, in his writings on the need for achievement, and by Maslow, in his writings on the need for self-esteem. Concepts of reasonableness and prudence provide (among reasonably matched people) a basis for a standard of excellence by which one can measure oneself. Since in the ordinary course of social interaction we know most intimately those people who are reasonably matched to us, our experience is that all people need is the same opportunities we have had.

Sociologically, one can observe that the concept of blame serves to support the status quo. Problems which in the liberal program might be seen as deriving from inequities in the basic structure of the current institutional arrangements, are seen in the conservative program as deriving from errors of the people experiencing the problems. These problems are accordingly seen as no occasion for change. The institutions of society are in military terms the high ground: it is extremely difficult to make social changes without command of these institutions. The institutions perform the functions of both communication and legitimation, two factors centrally involved in social change. By definition, it is the people who are able to succeed in our current social arrangements who operate and command these institutions, forming tightly knit social groups which can maintain and perpetuate the conservative project.

It is important to recognize that this last is not a conspiracy theory. The argument is not that there are secret meetings of the elite where they swear blood oaths to continue their oppression. Whether such meetings occur or not is scarcely relevant: the sociological basis of the conservative project is the existence of a widely shared perspective among the elite on the proper analysis of social issues. Secret meetings are unneeded when the mechanism of oppression is so well understood. Like Adam Smith's "invisible hand," no particular effort is needed (except continuing to act individually in accordance with the philosophy outlined above) to maintain the conservative project as it is.


To the four systems discussed above could be added many more. Let us note some of these in a more cursory fashion than above.

As has long been known, the structural-functionalist project in social science has the property of hermeticity: the existence of any social structure must be explained in terms of its function within the overall system, and given the necessity of such explanation, it is not surprising that one can be created, there being no limit to the efflorescence of such explanations. In the structural-functionalist system, our h0 is "The set of functions (f1,f2,...fn) explains the existence of all the structures in a society." This h0 can be challenged a posteriori by the request for an explanation of some particular structure. Obviously h0 is left intact if the structure can be explained. If the structure cannot be explained, a new function can be adduced and h0 modified appropriately: "This set of functions (f1,f2,...fn, and fn+1) explains the existence of all structures in a society." Lying behind this process, our H0 is simply the basic tenet of structural-functionalism: "All structures can be most fruitfully explained in terms of functions they perform." It is this that Levi-Strauss (1963:13) challenges when he tartly remarks, "... to say that a society functions is a truism; but to say everything in a society functions is an absurdity." Only boredom with the oppressive mass of explanation can bring the structural-functionalist project to a halt.

A closely related hermetic project is the field of belief systems studies, as has been pointed out by Rosenberg (1982). Each set of beliefs can be explained in terms of a set of explanatory variables (e.g., using factor analysis), but increases in the diversity of either the beliefs or the population studied yield only more factors rather than a closure or convergence of explanations. Again, the efflorescence of explanation halts only from boredom.

Finally, and again closely similar to structural- functionalism, are three examples from the field of positive literary criticism: the historical-biographical Marxist, and psychoanalytical methods. In all three methods the underlying assumption H0 is that a text can be best understood as the product of certain forces: the experiences of the author, in the historical-biographical method; the material interests of the author's class, in the Marxist method; or the author's unconscious drives, in the psychoanalytic method. This H0 results in a series of h0's of the form: "This text is understood as coming from this set of forces (f1,f2,...fn) in such-and-such a way." This h0 can be challenged by examples from the text and, if required, replaced with a new one, but the H0 remains.


With these examples in mind, we are now in a position to outline the characteristics of hermetic projects.

Figure 3: General Scheme

(1) Determine existence of problem.

(2) Narrow the question to one answerable in terms of a given paradigm.

(3) Change the model within the paradigm to a new model consistent with problem.

Note the psychological and sociological investment in the paradigm which facilitates the step from (1) to (2). Also note that steps (1) and (2) may not be especially distinct, in that the very denotation of a condition (i.e., set of circumstances) as a problem carries along with it (or, may carry along) a set of possible recourses: those within the project.

As shown in Figure 1, hermetic projects consist of three steps in a continuing cycle: first, an empirical fact is discovered which appears a posteriori to threaten the current system (represented by h0; second, the problem is narrowed to a request for a solution of a certain specified type, and it is this understanding of what modifications or changes are acceptable (represented by the a priori hypothesis H0) which constitutes the paradigm; and third, the required changes are made within the paradigm, with the result that a new h0 is established and the equilibrium of the paradigm is restored, beginning a new cycle. The critical step is the narrowing of the questions between steps (1) and (2). This narrowing is supported, as we have seen, by logical, psychological and sociological considerations. The logical considerations are of two forms: narrow and broad. Narrowly, the replacement of one h0 by another as constrained by H0 is perfectly logical, and it is this constraint, this cycle of replacement, which makes the project "hermetic." Broadly, H0 itself points to some truth about the situation. It cannot be said to be "logical" to believe H0, in that one cannot directly prove or disprove an a priori hypothesis, but it is at least "reasonable." Speaking generally, we can say that the psychological considerations stem from the great disruption involved in people's changing their world-view in the radical fashion implied by a paradigm shift. Again, at the same level of generality, we can say that the sociological considerations stem from the way in which social organization is created in the context of shared believe structures. To the extent that people depend on this organization to meet their needs in a continuing fashion, we can say that their concrete interests are served by maintenance of their (shared) belief structure H0.

8.1 The Limitations of Hermetic Projects

The issue, then is not that the paradigms are illogical. If the paradigm of Newtonian mechanics were capable of explaining all physical phenomena, presumably the scientific project would eventually establish the correct set of laws. Nor can we prove even now that Newtonian mechanics are wrong. In fact, there is every reason to believe that a Newtonian explanation of all known phenomena could be constructed, the only problem being (as it has been since relativity and quantum mechanics were first postulated) that the ratio of hypothesis to explanation seems to be very high compared to that of quantum mechanics. Relativistic quantum mechanics explains a lot with a little; Newtonian mechanics explains a lot with a lot. As I say, the problem is not that a given paradigm is unscientific. The issue is that the range of explanation is restricted. If the truth lies outside it, said truth cannot be discovered: in principle, the paradigm has the means neither to discover the truth nor to stop trying.

8.2 Overcoming Hermeticity

The problem we are faced with, then, is how one can prevent oneself from being caught inextricably in a hermetic project. Since hermetic projects have three components -- logic, psychology and sociology -- the answer must have three components also.

8.2.1 Logic

The logical answer is simply to distinguish between the forms of argumentation appropriate to the two levels of hypothesis. A posteriori falsification of h0 seems to be an adequate philosophical stance within the hermetic project. its logical status is well known, so I will deal with it no further. When one is arguing at the level of H0 -- that is, when one is arguing over the relative merits of alternative a priori (paradigmatic) positions, a different philosophical stance must be adopted. Such a stance is laid out in John Rawls A Theory of Justice, which terms the end state "reflective equilibrium -- a state in which, recognizing that intuition can modify moral principles and moral principles can modify intuition, both are in equilibrium, and which terms the mechanism for reaching such an equilibrium "the original position." Piaget's concepts of Assimilation, accommodation and equilibrium play the same part in his theory (Flavell 1963). Gandhi's technique of Satyagraha, according to Bondurant's (1971) exegesis, had the same function. That these solutions to a philosophical problem are so widespread testifies, I feel, to the fundamental nature of the problem posed by hermetic projects.

8.2.2 Psychology

[xx Realize that despair is o.k.? Embrace discourse?]

8.2.3 Sociology

[Abandon experimental success or failure as criteria of social value.]


Bauman, Carol (1980). A Conservative View of Women's Liberation. Reprinted in Rubenstein and Slife (eds) (1980:178 ff).

Blenkinsopp, Joseph (1971). Parrhesia: Free Speaking and Free Being in the Church. The Living Light 8(2, Summer):91-98.

Bondurant, Joan V. (1971). Conquest of Violence: The Gandhian Philosophy of Conflict. Berkeley: University of California Press.

Chilton, Stephen (1980a). Issues in the Application of an Alternate Income Definition. Washington D.C.: Congressional Research Service.

Chilton, Stephen (1980b). Issues in the Determination of Cash Value of Program Benefits from Survey Data. Washington, D.C.: Congressional Research Service.

Einstein, Albert (1905). Zur Elektrodynamik bewegter Körper [On the Electrodynamics of Moving Bodies]. xx

Flavell, John (19xx). The Developmental Psychology of Jean Piaget. xx:xx.

Ford, Daniel (1982). The Cult of the Atom. New Yorker 58(36):107 ff. & 58(37):45 ff.

Habermas, Jürgen (1983). Interpretive Social Science vs. Hermeneuticism. In Norma Haan et al., eds. Social Science as Moral Inquiry. New York: Columbia University Press.

Department of Health, Education and Welfare (19xx). The Measurement of Poverty. Washington, D.C.: U.S. Government Printing Office.

Holton, Gerald (1973). xx

Hewitt (1979). Self and Society (2nd ed.). Boston: Allyn & Bacon.

Hummel, Ralph and Robert Isaak (1980). Politics for Human Beings (2nd ed.) Belmont, CA: Duxbury Press.

Jernigan, Kenneth (1982). [New Technology for the Blind]. Family Weekly (December 5, 1980, p.2).

Kuhn, Thomas (1970). The Structure of Scientific Revolutions (2nd ed.) Chicago: University of Chicago Press.

Levi-Strauss, Claude (1963). Structural Anthropology (translated by Claire Jacobson and Brook Grundfest Schoepf). New York: Basic Books.

Mach, Ernst (1883). The Science of Mechanics [See Holton] xx.

Maslow, Abraham (19xx). Motivation and Personality. xx:xx.

McClelland, David (1976). The Achieving Society New York: Irvington Publishers, Inc.

Paheau, Françoise (1980). Narrative Semeiotics in France: A Critical Explication Unpublished Ph.D. dissertation, University of Utah.

Rawls, John (1971). A Theory of Justice. Cambridge: Belknap.

Rubenstein, Joseph and Brent D. Slife, eds. (1980). Taking Sides: Clashing Views on Controversial Psychological Issues. Guilford, CT: Dushkin Publishing Group, Inc.

Ryan, William (1976). Blaming the Victim. New York: Random House.

Suppe, Frederick, ed. (1977). The Structure of Scientific Theories. Urbana: University of Illinois Press.

Teague, Kathy (1980). The Most Admired Women. Reprinted in Rubenstein and Slife (eds) (1980:178 ff).

Page URL: http://www.d.umn.edu/~schilton/Articles/Hermetic.html
Page Author: Stephen Chilton
Link to Home Page: www.d.umn.edu/~schilton/index.html
Last Modified: January 21, 2002
Mail suggestions and comments, especially re. typos and other errors, to: schilton@d.umn.edu
Honor Roll of Proofreaders and Colleagues: www.d.umn.edu/~schilton/Honor.html
This way to the UMD homepage
This way to the Political Science Department home page

The University of Minnesota is an equal opportunity educator and employer.
Copyright © 1999 Regents of the University of Minnesota. All rights reserved.