AnthonyFlood.com
Panentheism. Revisionism. Anarchocapitalism.
Process, Insight, and Empirical Method
An Argument for the Compatibility of the Philosophies of Alfred North Whitehead and Bernard J. F. Lonergan and Its Implications for Foundational Theology.
A Dissertation Submitted to the Faculty of the Divinity School, The University of Chicago, for the Degree of Doctor of Philosophy
December 1983
Thomas Hosinski, C.S.C.
Chapter I:
Whitehead’s and Lonergan’s Interpretations of Empirical Scientific Method and Philosophic Method [continued]
Lonergan’s Interpretation of Scientific and Philosophic Method
The Method of Empirical Science
If the problem confronting us in describing Whitehead’s analysis of empirical scientific method was the lack of a single sustained analysis of scientific method in Whitehead’s writings, the opposite problem confronts us when describing Lonergan’s analysis. He devotes the first five chapters of Insight [Ibid., pp. 3-172.] to a thorough, closely reasoned analysis of scientific method, and if I am to keep this study within reasonable limits, I can do little more than provide a summary of the major features of his interpretation. I will begin this section with a brief discussion of Lonergan’s summary statement of his interpretation, and then proceed to fill in that summary view by reference to the detailed analysis in Insight.
The Summary View
The clearest, most easily understood summary of Lonergan’s interrpretation of empirical scientific method occurs in the first chapter of Method in Theology.
Method, pp. 4-6. The first chapter of Method attempts to summarize those conclusions of Insight necessary for basing Lonergan’s interpretation of method in theology.
. . . in the natural sciences method inculcates a spirit of inquiry and inquiries recur. It insists on accurate observation and description: both observations and descriptions recur. Above all it praises discovery and discoveries recur. It demands the formulation of discoveries in hypotheses, and hypotheses recur. It requires the deduction of the implications of hypotheses, and deductions recur. It keeps urging that experiments be devised and performed to check the implications of hypotheses against observable fact, and such processes of experimentation recur.
These distinct. and recurrent operations are related. Inquiry transforms mere experiencing into the scrutiny of observation. What is observed is pinned down by description. Contrasting descriptions give rise to problems, and problems are solved by discoveries. What is discovered is expressed in a hypothesis. From the hypothesis are deduced its implications, and these suggest experiments to be performed. So the many operations are related; the relations form a pattern; and the pattern defines the right way of going about a scientific investigation.
Finally, the results of investigations are cumulative and progressive. For the process of experimentation yields new data, new observations, new descriptions that may or may not confirm the hypothesis that is being tested. In so far as they are confirmatory, they reveal that the investigation is not altogether on the wrong track. In so far as they are not confirmatory, they lead to a modification of the hypothesis and, in the limit, to new discovery, new hypothesis, new deduction, and new experiments. The wheel of method not only turns but also rolls along. The field of observed data keeps broadening. New discoveries are added to old. New hypotheses and theories express not only the new insights but also all that was valid in the old, to give method its cumulative character and, to engender the conviction that, however remote may still be the goal of the complete explanation of all phenomena, at least we now are nearer to it than we were. [Ibid., pp. 4-5.]
This passage gives us the basic framework of Lonergan’s interpretation of empirical scientific method, and some attention to its major features will greatly facilitate our later discussion.
First, there are a few general points to be noted about empirical scientific method as an activity. It is clear that Lonergan interprets that method as an activity composed of several distinct operations which recur, and recur in a definite pattern of relationship. Hence empirical scientific method has a structure, but it is a dynamic structure of distinct operations recurring in a pattern. Further, this pattern of recurrent operations is normative: it defines the proper way to conduct the activity of scientific inquiry. I shall have more to say about these general characteristics of empirical scientific method later; for now it is enough to note them.
Lonergan interprets empirical scientific method as having four major stages or moments
Although I will discuss the four major moments of empirical scientific method in the context of the quotation from ibid., pp. 4-5, I might note in passing Lonergan’s own summary of these elements, Insight, p. 79: “On the previous analysis, then, empirical method involves four distinct elements, namely
(1) the observation of data,
(2) insight into data,
(3) the formulation of the insight or set of insights, and
(4) the verification of the formulation.”
The reader will note that this four-part structure does not exactly correspond to what Lonergan will later describe as the three levels of cognitional process (Insight, p. 274). Lonergan there lists insights and formulations as two elements of the second level of cognitional process. See Thesis, p. 84.
which are connected by movements or activities of the inquiring mind. It will facilitate the discussion of this dynamic structure if I anticipate my later observations to the extent of noting that in Lonergan’s analysis scientific inquiry is basically the activity of raising and answering questions on succeedingly higher levels of consciousness. The first moment in empirical scientific method, then, can be defined as intelligence questioning experience: what is it that I am seeing, hearing, tasting, smelling, touching, feeling? The scientist’s first step in answering such questions is to scrutinize what is being experienced, to observe, and to describe what is being observed. Obviously, observation and description, while both operations of the first major moment of empirical method, are not the same operation. Observation would have little to guide it were it not for the requirement that what is being observed be described. The necessity of observation resulting in a description forces observation to a first level of acuteness. As Lonergan says in the above quotation, “What is observed, is pinned down by description.”
Questioning, however, does not cease with description. Description alone does not explain the data, and contrasting descriptions of the data to be explained raise problems for understanding. The attempt to resolve such problems constitutes the second major moment of empirical method, which Lonergan calls “discovery,” or more frequently, “insight.” In this moment an understanding has been reached, an answer discovered to the questions raised following the observation and description of the data. The mind has been confronted with a problem, and the second major moment of empirical method is the sudden resolution of that problem, opening the way to understanding.
Lonergan, of course, has a good deal to say about the elements of insight, and I will discuss this shortly.
But if insight is a grasp of the resolution of a problem, if it is understanding in its initial phase, insight alone is not yet the expression of understanding, nor is it explanation. The insight or set of insights must be expressed or, more technically, formulated in an hypothesis. The hypothesis states in a more generalized form and in precisely defined language what the understanding grasped in the insight is. The hypothesis is, in short, the formalized conceptualization of the insight. The insight: is, by its nature, a private event that is in an important sense preconceptual.
See Insight, p. 59: “. . . an insight is neither a definition nor a postulate nor an argument but a preconceptual event.”
If it is not to remain private, it must be formulated in precise language, public concepts, and in a general form so that the understanding it has grasped can be communicated to those who have not had the insight. In the work of the empirical sciences, this is accomplished by the formulation of hypotheses. The formulation of hypotheses, then, is really the formulation of the understanding gained in the insight, and this constitutes the third major moment or stage in the dynamic of empirical scientific method. This third moment is connected to the second moment (insight) by a recurrence of questioning in the inquirer. How does this insight resolve the problem? How does it explain the data?
Perhaps an example would be of assistance in explaining the connection between the second moment of insight and the third moment of hypothesis formation. Consider Archimedes sitting in the baths of Syracuse, pondering the problem with which King Hiero had presented him: how to discover whether a gold crown the king had commissioned was made of pure gold or not.
This is the example Lonergan uses to discuss the elements of insight. See ibid., pp. 3-6.
Archimedes has his insight and rushes naked from the baths shouting “Eureka!” He has grasped how to resolve this problem. But had he not been able to formulate his understanding in an explanatory hypothesis so as to communicate it to the king, Archimedes might forever have been considered an eccentric. He had to be able to answer the king’s questions, and his own. He had to be able to explain how the insight would solve the problem. It is because Archimedes was able to formulate his insight in the principles of hydrostatics later known as specific gravity and displacement that this particular incident in Syracuse is a lasting part of the history of science. Archimedes’ private insight became, with his formulated explanation, a part of human understanding.
But even the precise formulation of the understanding in an hypothesis does not terminate the questioning of scientific inquiry. Further questions arise: Is this understanding correct? Is it so? Is it true? In order to answer these questions the inquirer proceeds to operations which connect the third moment of hypothesis formation with the fourth moment of testing. These operations are deducing the implications of the hypothesis and forecasting what will happen in certain circumstances according to the understanding embodied in the hypothesis. These deductions and forecasts suggest experiments that can be carried out in order to test the understanding formulated in the hypothesis. Thus the inquirer moves toward the fourth major moment of empirical scientific method: testing.
In this fourth major moment the hypothetical understanding is tested by the performance of the suggested experiments. The results of the experiments produce new data, new observations, and new descriptions which lead the scientist to make a judgment
In the summary passage I am commenting upon, Lonergan does not use the word “judgment” to describe the conclusion of the fourth moment of scientific method. However, as is abundantly clear from his other discussions, judgment is the operation involved at this moment. See Method, pp. 6-10 and Insight, pp. 271-278.
concerning the hypothetical understanding: the experimental results either confirm the hypothesis or they do not confirm it (or, we might add, the results are inconclusive and further testing is required). It is important to note the way in which Lonergan describes cases in which the hypothesis is confirmed or verified. In Lonergan’s analysis such confirmation is a limited form of verification: “In so far as they are confirmatory, they reveal that the investigation is not altogether on the wrong track.” There is no question here of absolute certainty; the judgments are more or less probable, always remaining open to modification by further data, insights, understandings, and judgments. In cases where the experimental results lead the scientist to judge that the hypothesis has not been confirmed, they cause the scientist to modify the hypothesis or, in cases in which the hypothesis has been judged to be completely falsified, to abandon the hypothesis and search for “new discovery, new hypotheses, new deduction, and new experiments.” In short, the result of the scientist’s judgment in all cases is to continue the inquiry by a repetition of all the operations, though now guided and informed by what has been learned in the course of the inquiry.
Hence Lonergan describes the results of the application of empirical scientific method as “cumulative and progressive.” It is cumulative because, as the field of observed data is broadened and as new discoveries are made beyond the old, “new hypotheses and theories express not only the new insights but also all that was valid in the old . . . .” It is progressive because scientists, no matter how deeply aware they might be of the breadth of our ignorance, are convinced that they are now at least a little nearer the truth than before. It is in this connection that Lonergan uses an image, a metaphor, that is striking in its ability to communicate his interpretation of empirical scientific method. Lonergan says, “the wheel of method not only turns but also rolls along.” The activity of method in science is like a wheel. Empirical method is a structured, patterned thing, and the operations which constitute the elements of that patterned activity continually recur, each operation depending on all that have gone before. But this structured pattern of recurrent operations does not merely spin in the same place, with a deadening lack of advancement, like some wheel lifted from the ground; it moves creatively toward the goal of knowledge, rotating on the axle of inquiry, a relentless questioning that will not cease until there are no further questions to be asked, driven by the eros of the human mind to know.
Finally, this method is normative. This pattern of related and recurrent operations is understood by scientists to be the right way of conducting a scientific investigation. An inquiry that is not conducted according to this pattern is not regarded as scientific, nor will the community of scientists accept any alleged “discoveries” that have not been formulated and tested in accord with the pattern of empirical scientific method.
One important element of empirical scientific method Lonergan does not allude to in this summary is the communal dimension of science as expressed in its demand for repeatability of performance.
There is one further point to be noted about Lonergan’s summary description of empirical scientific method. Lonergan is not describing some specific method of the empirical sciences. Rather, he is describing the general method of questioning the scientist uses as he or she pursues an inquiry, the general method of thought that governs the development, use, and application of all special methods in the particular sciences.
After the passage quoted at the beginning of this section (Thesis, p. 56-57), Lonergan goes on to say, “Such, very summarily, is method in the natural sciences. The account is far indeed from being sufficiently detailed to guide the natural scientist in his work. At the same time it is too specific to be transposed to other disciplines.” Method, p. 5.
This is not some theoretical model of scientific method drawn up to fit a prior philosophical theory; it is intended to be a description of how scientists actually proceed. It is intended to describe the patterned structure of operations they actually follow in the conduct of their inquiries.
Finally, Lonergan points out an important characteristic of the opperations which make up the pattern of empirical scientific method. These operations are of two different sorts: “logical” and “non-logical.” [Method, p. 6.] “Logical” operations are operations on propositions, terms, and relationnships. “Non-logical” operations are operations of inquiry, observation, discovery, experiment, synthesis, and verification. Lonergan points out that
modern science derives its distinctive character from this grouping together of logical and non-logical operations. The logical tend to consolidate what has been achieved. The non-logical keep all achievement open to further advance. The conjunction of the two results in an open, ongoing, progressive, and cumulative process. [Ibid.]
The importance of the non-logical operations will be discussed in later sections of my study. For now it is enough to note that they are present as integral elements of empirical scientific method.
The Detailed Discussion: The Analysis in Insight
My discussions in this section will not bear an immediate resemblance to the summary discussion I have just concluded. This is because I shall be following the plan of Insight, which begins with a relatively static account of the elements involved in an insight, and gradually moves on to higher viewpoints, considering the method of empirical science in all its dynamism and its activities. It is necessary to follow the plan of Insight and begin with the relatively static account because in the first chapter, “Elements,” Lonergan lays the groundwork for his entire analysis by describing the basic characteristics of insights and the conditions that stimulate their occurrence. Before beginning my discussion, and at the risk of sounding apologetic, I must once again stress that if I am to keep this chapter within manageable bounds, I can only summarize and highlight the principal points of Lonergan’s minutely detailed study.
“Elements”: Chapter I of Insight
Using the example of Archimedes’ discovery, Lonergan begins by noting that insight has five characteristics. Insight
(1) comes as a release to the tension of inquiry,
(2) comes suddenly and unexpectedly,
(3) is a function not of outer circumstances but inner conditions,
(4) pivots between the concrete and the abstract, and
(5) passes into the habitual texture of one’s mind. [Insight, pp. 3-4.]
For my purposes, just a few comments on characteristics (1), (3), and (4) are required. First, it is to be noted that insights come in the context of the tension of inquiry. This indicates the absolutely central role played by questioning, and I will return to this point repeatedly. Characteristic (3) points out that insights arise because of the internal conditions of the inquirer, not because of external circumstances. This is not to say that external circumstances are not involved in or cannot spark the occurrence of an insight. Rather, Lonergan is pointing out that no matter what the external conditions might be, the insight will not occur if the proper internal conditions are not present in the inquirer. Among these conditions is “the accurate presentation of definite problems” [Ibid., p. 5.]; that is, the questioning driving the inquiry is not aimless, but concerns specific and well-formulated problems. Characteristic (4) is a very important observation on the nature of an insight: it pivots between the concrete and the abstract. Insights occur because concrete problems are being addressed, and they are valuable because they can be concretely applied.
But because the significance and relevance of insight goes beyond concrete problem or application, men formulate abstract sciences with their numbers and symbols, their technical terms and formulae, their definitions, postulates, and deductions. Thus, by its very nature, insight is the mediator, the hinge, the pivot. It is insight into the concrete world of sense and imagination. Yet what is known by insight, what insight adds to sensible and imagined presentations, finds its adequate expression only in the abstract and recondite formulations of the sciences. [Ibid., p. 6.]
These considerations lead us directly to the second major movement of inquiry: the urge, desire, even necessity of expressing the insight exactly, formulating it in technical language so that its meaning is clear. In discussing how an insight struggles for exact expression, Lonergan notes the important roles played by images, properly framed (limited) questions and experimentation with concepts in order to answer the question. [Ibid., pp. 7-9.] When the insight is achieved and issues in a definition (to consider the simplest case),
The answer is a patterned set of concepts. The image strains to approximate to the concept. The concepts, by added conceptual determinations, can express their differences from the merely approximate image. The pivot between images and concepts is the insight. And setting the standard which insight, images, and concepts must meet is the question, the desire to know, what could have kept the process in motion by further queries, had its requirements not been satisfied. [Ibid., p. 10.]
This analysis not only reveals that there is a dynamic involved in the exact expression of an insight, it also allows us to discover the root of the normative nature of empirical scientific method in the nature of questioning itself. The motive force of the activity which is scientific investigation is questioning, driven in turn by the desire to know, and the pattern of empirical method is normative because only by asking and answering the questions in that way are the demands of the questions satisfied. Thus the norm lies in the dynamic of inquiry itself, and in turn is dependent upon the unrestricted desire to know which is the eros of the human inquirer. I shall return to this point in a later section, but it is important to note that one can detect the normative role of questioning not only in the larger dynamic of empirical scientific method, but also in the basic elements of insight.
Lonergan goes on, in a third major point, to note that insights, once they have been expressed in exact and technical language, do not remain in isolation. Rather, the dynamic of questioning continues and brings insights together in groups. Inquiry continues in succeedingly higher viewpoints. [Ibid., pp. 13-17.] In discussing the meaning, occurrence, and necessity of higher viewpoints, Lonergan notes the importance of symbolism
[Ibid., pp. 17-19. The “symbolism” being discussed is mathematical symbolism.]
in the dynamic of inquiry moving to higher viewpoints. For my purposes, there are two characteristics of symbolism that ought to be mentioned. First, symbolism in mathematics constitutes a heuristic technique. The nature of the symbolism names the unknown that is being sought, and the rules governing the manipulation of the symbolism draw the inquirer on toward the solution of the problem, toward the discovery of the unknown. But mathematical symbolism is heuristic in another way as well: it offers clues, hints, and suggestions that provide the image in which the inquirer may grasp by insight the rules for the next stage of mathematical development (a higher viewpoint).
See Lonergan’s discussion, ibid., pp. 18-19.
As we shall shortly see, empirical scientific method is itself such a heuristic technique guiding the scientist on toward the goal of discovery.
A fourth important observation Lonergan makes is that there are two different kinds of insights. The kind I have been discussing thus far he names “direct” insights. He now observes that there is also a kind of insight he names “inverse.” [Ibid., pp. 19-25.] A direct insight discovers the solution to a problem or the answer to a question; as an act of understanding it finds an expected intelligibility in the data it ponders. An inverse insight discovers that the problem has no solution, that this question has no answer; it denies the expected intelligibility in the data it ponders.
. . . while the conceptual formulation of direct insight affirms a positive intelligibility though it may deny expected empirical elements, the conceptual formulation of an inverse insight affirms empirical elements only to deny an expected intelligibility. [Ibid., p. 19.]
In short, an inverse insight is an act of understanding that on this question, regarding this empirical data, intelligibility cannot be discovered. As will become apparent later in my discussion, this description of inverse insight is important in Lonergan’s analysis because it serves as the basis for grounding statistical method as a valid part of empirical scientific method. [Ibid., pp. 53-58.]
In exploring the connection of inverse insights with ideas, principles, methods, and techniques of “quite exceptional significance,” Lonergan introduces the final interpretive “element” in his initial analysis of insight, the notion of an “empirical residue.” [Ibid., pp. 25-32.] The empirical residue
(1) consists in positive empirical data,
(2) is to be denied any immanent intelligibility of its own, and
(3) is connected with some compensating higher intelligibility of notable importance. [Ibid., pp. 23-26.]
Thus the elements in the empirical residue are experienced, observed, described, conceived, named, thought about, and affirmed or denied; but they are not objects of a direct insight, and so they cannot be explained. The clearest examples of elements pertaining to the empirical residue are particular places and particular tunes. These are positive aspects of our experience, and we note that each differs from every other as a matter of fact. Yet “there is no immanent intelligibility to be grasped by direct insight into that fact.” [Ibid., p. 26.] A further consideration is that “because the differences of particular places and particular times involve no immanent intelligibility of their own, they do not involve any modification in the intelligibility of anything else.” [Ibid., pp. 27.] The importance of this consideration is that it is the ground of the possibility of scientific collaboration and also of scientific generalization. [Ibid., pp. 28-30.] It is precisely because the differences of particular places and times do not modify the intelligibility of anything else that scientists in different places and times are able to pool their results and depend on each others’ work. Even more importantly, as particular times and places differ, so do particular individuals within a class differ; “but the ultimate difference in our universe is a matter of fact to which there corresponds nothing to be grasped by direct insight.” Thus even though there are differences between every individual in a given class, scientists are able to generalize about the class. For example, every hydrogen atom differs from every other hydrogen atom in some empirical respects, yet scientists are able to speak of hydrogen atoms in general without being required to explain the uniqueness of each particular hydrogen atom.
This leads immediately to the closely related topic of abstraction. [Ibid., pp. 30-31. See also pp. 88-89 for additional discussion.] Abstraction, Lonergan observes, is the “selectivity of intelligence,” and “to abstract is to grasp the essential and to disregard the incidental, to see what is significant and set aside the irrelevant, to recognize the important as important and the negligible as negligible. [Ibid., p. 30.] What is essential, significant, and important is “the set of aspects in the data necessary for the occurrence of the insight or insights” and the related concepts “necessary for the expression of the insight or insights.” What is incidental, irrelevant, and negligible is the other “aspects of the data that do not fall under the insight or insights” as well as the concepts that correspond to these aspects. In a science or group of allied sciences that reaches full development, Lonergan contends, “the incidental, irrelevant, negligible consists in the empirical residue that, since it possesses no immanent intelligibility of its own, is left over without explanation . . . .”
Ibid., p. 31. Even if one is to take a science or group of allied sciences that “reaches full development” as a limit case that science approaches but never realizes, it seems to me that there is something missing from Lonergan’s account of abstraction, namely, the observation that a particular purpose governs the making of all abstractions. Something is irrelevant or negligible only given a certain purpose. Thus, in consigning all that is not relevant to a “fully developed” science to the empirical residue (by definition without immanent intelligibility), is not Lonergan overlooking a fundamental feature of abstraction? For example, if Whitehead’s account of abstraction is accurate, then it seems that scientific abstractions by their nature abstract not only from the empirical residue, but also from other aspects of events that do possess immanent intelligibility when approached by inquiry with purposes different than the scientific. This is a question to which I will have to return in a later part of my study.
The importance for my purposes of this analysis which Lonergan gives under the heading of “the empirical residue” perhaps ought to be made explicit. First the notion of the empirical residue will be a significant factor in Lonergan’s analysis of statistical method. But more importantly for my purposes, it is clear even at this stage that abstraction and generalization are the two activities involved in the practice known as induction. The analysis of the notion of the empirical residue and how it is related to the activities of abstraction and generalization will provide the ground upon which Lonergan will attempt to exhibit the validity of the practice of induction, arguing that both abstraction and generalization (which, obviously, depends on abstraction) are grounded and find their validity in the very dynamism of inquiry itself.
The Heuristic Structures of Empirical Method: Chapter II of Insight
Lonergan now moves his attention to the dynamic method of the empirical sciences, and he begins by noting the similarities and dissimilarities between mathematical and scientific insights.
Ibid., pp. 33-35. I will restrict my discussion of Chapter II to only those points important for my purposes.
There are circuits of development in both mathematical and scientific method, Lonergan notes, but the circuits are of different sorts. Mathematics reaches higher viewpoints by following a circuit that can be outlined as follows: initial images give, rise to insights; insights give rise to definitions and postulates; definitions and postulates guide the use of the heuristic device of symbolic operations; and symbolic operations give rise to “a more general image in which the insights of the higher viewpoint are emergent” [Ibid., p. 35.] The circuit of scientific inquiry differs in this way:
The operations that follow upon the formulation of laws are not merely symbolic. For the formulation expresses a grasp of possibility. It is a hypothesis. It provides a basis for deductions and calculations no less than mathematical premises. But it also provides a basis for further observations and experiments. It is such observation and experimentation, directed by a hypothesis, that sooner or later turns attention to data that initially were overlooked or neglected; it is attention to such further data that forces the revision of initial viewpoints and effects the development of empirical science. [Ibid.]
Thus the circuit of empirical scientific method is similar in structure to mathematical method, but is distinguished because it must be empirical; it must attend to its data, which are external to its own hypothetical formulations. The movement to a higher viewpoint occurs as a result of comparing the hypothetical formulation and its deduced implications to further data by means of experimentation and renewed observation.
Lonergan summarizes all this even more briefly: “The circuit, then, of mathematical development may be named immanent; it moves from images through insights and conceptions to the production of symbolic images whence higher insights arise. But the circuit of scientific development includes action upon external things; it moves from observation and experiment to tabulations and graphs, from these to insights and formulations, from formulations to forecasts, from forecasts to operations, in which it obtains fresh evidence either for the confirmation or for the revision of existing views.” Ibid.
Lonergan now attempts to answer a puzzling question.
Scientists achieve understanding, but they do so only at the end of an inquiry. Moreover, their inquiry is methodical, and method consists in ordering means to achieve an end. But how can means be ordered to an end when the end is knowledge and the knowledge is not yet acquired?
Ibid., p. 44. This quotation is taken from a summary discussion. Lonergan pursues this question on pp. 35-46.
In short, how is it possible for a scientist even to begin an investigation when he or she does not yet know what it is that he or she is trying to know? “The answer to this puzzle is the heuristic structure. Name the unknown. Work out its properties. Use the properties to direct, order, guide the inquiry.” [Ibid., p. 44.] And so just as the mathematician is able to devise heuristic techniques to guide inquiry, [See ibid., pp. 17-19, and Thesis, pp. 65.] the scientist, too, is able to devise and use an heuristic structure to guide inquiry. Lonergan gives a good summary of what the heuristic structure in empirical scientific method is:
What is to be known inasmuch as data are understood is some correlation or function that states universally the relations of things not to our senses but to one another. Hence, the scientific anticipation is of some unspecified correlation to be specified, some indeterminate function to be determined; and now the task of specifying or determining is carried out by measuring, by tabulating measurements, by reaching an insight into the tabulated measurements, and by expressing that insight through some general correlation or function that, if verified, will define a limit on which converge the relations between all subsequent appropriate measurements.
Insight, p. 44. See ibid., pp. 44-45 for Lonergan’s summary of the further “enrichments” of this basic anticipation and procedure. One that perhaps I ought to note in passing is that science anticipates the independence of the general correlation or function from the empirical residue, and this is expressed in physics as “the invariance of principles and laws under groups of transformations.” See also ibid., pp. 39-43.
As the reader will no doubt note, this “heuristic structure” is a more specific application of the general dynamic structure of empirical method I discussed in the preceding subsection. [Thesis, pp. 56-62.] The problem which gives rise to inquiry is here “the anticipation of some unspecified correlation.” Observation and description are here “measuring” and “tabulating measurements.” The insight grasping the solution of the problem is here “an insight into the tabulated measurements.” The expression or formulation of the insight in an hypothesis is here “expressing that insight through some general correlation or function,” and, as hypothesis attempts to express what will be the case for all similar phenomena, here the “general correlation or function” attempts to “define a limit on which converge the relations between all subsequent appropriate measurements.” Finally, just as hypotheses must be verified (that is, tested), so here too the general correlation or function must be verified. Hence, while Lonergan is here discussing a rather specific procedure within the physical sciences, this analysis also reveals that empirical method itself is a heuristic structure. Nor ought it to be overlooked that, even at this early stage in his study, Lonergan detects the same basic heuristic structure in “pre-scientific” thought as well. [See Insight, p. 44.] Though my concern at this point in my study is with empirical scientific method alone, the last-mentioned point will prove to be of some import in my later analysis.
Lonergan names the heuristic structure just described “classical,” and proceeds to a lengthy analysis of a second “heuristic structure,” the statistical. This analysis is designed to exhibit why statistical science cannot be dismissed as “a mere cloak for ignorance.” His purpose is
not to work out definitive foundations for statistical science but to grasp in some fashion the statistical heuristic structure that not only tackles specific problems but also develops its own methods as it goes along and thereby sets up an exigence for a succession of new and better foundations.
Ibid., p. 53. The analysis of statistical method, including distinguishing between systematic and non-systematic processes, occupies pp. 46-68.
To this end, Lonergan discusses the differences between systematic and non-systematic processes,
Ibid., pp. 47-53. For a critical analysis of the intelligibility of this distinction, see Harold H. Kuester, “The Epistemology of Michael Polanyi: A Solution to Current Epistemological Difficulties,” 2 vols. (Ph.D. dissertation, The Divinity School, University of Chicago, 1975), 2: 494-515.
and notes that while classical heuristic structure studies systematic and ignores non-systematic processes, statistical heuristic structure makes non-systematic processes its field of inquiry. The “radical” difference in mentality between classical and statistical inquirers is to be explained as based upon “something like an inverse insight,” [Insight, p. 54.] which denies intelligibility to random differences that occur in frequency patterns. [Ibid., pp. 54-55.] The statistical inquirer, in a positive insight, affirms intelligibility in what classical heuristic structure neglects, [Ibid., pp. 56-58.] and denies this intelligibility to random differences from the average frequency. [Ibid., p. 58.] This intelligibility is expressed in the concept of probability. [Ibid., pp. 58-62.] Lonergan goes on to discuss the insight which is at the root of the concept of probability.
By that insight the inquirer abstracts from the randomness in frequencies to discover regularities that are expressed in constant proper fractions named probabilities. There results the solution of two outstanding methodological problems. Because the probabilities are to hold universally, there is solved the problem of reaching general knowledge of events in non-systematic processes. Because states are defined by the association of classes of events with corresponding probabilities, there is by-passed the problem of distinguishing and listing non-systematic processes. However both the probabilities and the states they define are merely the fruits of insight. They are hypothetical entities whose existence has to be verified and, in fact, becomes verified in the measure that subsequent frequencies of events conform to probable expectations. In turn, this need of verification provides a simple formulation for the notion of a representative sample. For a set of relative actual frequencies is a representative sample if the probabilities to which they lead prove to be correct. . . . It follows that the basic practical problem of statistical inquiry is the selection of representative samples and, indeed, that its solution must depend not merely on a full theoretical development of statistical method but also on the general knowledge of individual investigators and on their insights into whatever specific issues they happen to be investigating.
Lonergan establishes this parallelism of dynamic structure in a more detailed analysis, ibid., pp. 63-66. Especially important is his conclusion that while the need for verification is the same in both classical and statistical heuristic structure, verification does not have the same meaning in both. See ibid., pp. 65-66.
This passage is important for several reasons. First it establishes that statistical inquiry does follow the basic structure of empirical method: it has insights, it expresses those insights in a hypothetical formulation, and this hypothetical formulation needs to be verified. But it also introduces a point which will become quite important later in my study. As Lonergan says, the need for verification of statistical hypotheses makes the selection of representative samples the basic practical problem, and the solution of this problem depends in part on the “general knowledge of individual investigators and on their insights . . .” In other words, the making of probable judgments of verification
See ibid., pp. 67-68 on the important and crucial distinction to be made between probable judgments of verification and judgments of the probable occurrence of events.
depends in part on the knowledge of the investigator, and so there is an undeniably personal element to the knowledge arrived at in making such judgments of verification.
I might note that Lonergan here expresses a position quite similar to that of Michael Polanyi. I will discuss this and other similarities to Polanyi in Chapter II.
Or to put it another way, in order to understand how it is that inquirers pursue understanding and come to knowledge through the making of probable judgments, one must be more concerned with the knower than with what is known. This points out the direction that Lonergan’s study follows. As he says at the end of the chapter on the heuristic structures of empirical method,
Our goal is the concrete, individual existing subject that intelligently generates and critically evaluates and progressively revises every scientific object, every incautious statement, every rigorously logical resting place that offers prematurely a home for the restless dynamism of human understanding. Our ambition is to reach neither the known nor the knowable but the knower. [Insight, p. 69.]
The Canons of Empirical Method: Chapter III of Insight
Returning to empirical scientific method, we may note that Lonergan has established the existence of two sorts of heuristic structures, the 4 classical and the statistical,
It might be helpful to note one of Lonergan’s later summaries of these two sorts of heuristic structures. “A classical heuristic structure is intelligent anticipation of the systematic-and-abstract on which the concrete converges. A statistical heuristic structure is intelligent anticipation of the systematic-and-abstract as setting a boundary or norm from which the concrete cannot systematically diverge.” [Ibid., p. 103.]
and he has argued that they are parallel in structure. He later notes,
Of themselves, heuristic structures are empty. They anticipate a form that is to be filled. Now just as the form can be anticipated in its general properties, so also can the process of filling be anticipated in its general properties. There exist, then, canons of empirical method. [Ibid., p. 103.]
These canons of empirical method govern the way in which an inquiry in empirical science is to be conducted, and Lonergan’s “single purpose is to reveal the intelligible unity that underlies and accounts for the diverse and apparently disconnected rules of empirical method.” [Ibid., p. 71.] My discussion of these canons is not intended to be a complete summary of Lonergan’s analysis of the canons, since I will have occasion to return to some of the issues raised in his discussions in later chapters of my study. For now I am interested simply in gaining a general understanding of the meaning of these canons and in noting a few areas of import for my purposes.
Lonergan states and discusses six canons of empirical scientific method: selection, operations, relevance, parsimony, complete explanation, and statistical residues.
Ibid., p. 70. Pp. 71-102 contain the detailed discussions of the canons.
The canon of selection states that the inquiry of empirical science is restricted to data of sensible experience. On the negative side it informs the empirical scientist to avoid questions that cannot be settled by observation and experiment; on the positive side “it directs the scientist’s efforts to the issue that he can settle by the decisive evidence of observation and experiment.”
Ibid., p. 72.
In discussing this canon, Lonergan makes three points to which I want to draw attention because of their importance later in the study. First, he notes that the canon of selection does not deny intelligibility to all other questions besides empirical scientific ones.
Questions that do not satisfy the canon of selection do not arise within the confines of empirical science, but it does not follow immediately that they do not arise at all. Issues that cannot be settled by observation or experiment cannot be settled by empirical method, but it does not follow immediately that they cannot be settled at all. [Ibid.]
Secondly, Lonergan entertains the notion that empirical method might be applied, at least in its essential features, to the data of consciousness as well as the data of sense. If it can be, he decides, then such a method would be named “a generalized empirical method.” [Ibid.] This question will not be taken up for some time in Lonergan’s study; but it is important to note where it originally arises. Finally, returning more immediately to the canon of selection, Lonergan comments on the fact that scientific observation is not some passive recording of sense impressions, but rather demands the full and vital participation of the scientist. The scientist as subject must be involved, must participate, if scientific observation is to take place: “it is not by sinking into some inert passivity but by positive effort and rigorous training that a man becomes a master of the difficult art of scientific observation.” [Ibid., p. 74; see also p. 73.] We will later see the full importance of this fact that scientific inquiry demands the full and active participation of the scientist.
The canon of operations describes how the activities or operations of science are guided by the laws which are the formulated insights arrived at in scientific inquiry.
. . . the laws provide premises and rules for the guidance of human activity upon sensible objects. Such activity, in its turn, brings about sensible change to bring to light fresh data, raise new questions, stimulate further insights, and so generate the revision or confirmation of existing laws and in due course the discovery of new laws. [Ibid., p. 74.]
Lonergan goes on to discuss how the canon of operations is a principle of cumulative expansion, of construction, of analysis, of cumulative verification, how it provides a test of the impartiality and accuracy of observations, how it is a principle of systematization, and how it is a source of higher viewpoints. [Ibid., pp. 74-76.]
The canon of relevance states the type of understanding proper to empirical science; “it states that empirical inquiry primarily aims at reaching the intelligibility immanent in the immediate data of sense.” [Ibid., pp. 76, 77.] This intelligibility that is immanent in the data of sense, according to the canon of relevance, resides in the relation of things to each other, not in their relation to our senses, and it consists not in an absolute necessity, but in a realized possibility. [See ibid., pp. 77-78.] To use Aristotelian vocabulary, the canon of relevance states that the intelligibility scientific inquiry is seeking is neither final, material, instrumental, nor efficient causality, but rather formal causality.
See ibid., pp. 76-78. See p. 78 on the possible misinterpretations of this term that are to be excluded.
One point to note is that the canon of relevance reveals the “two distinct grounds” upon which empirical science rests: “As insight grasping the possibility, it is science. As verification selecting the possibilities that are in fact realized, it is empirical.” [Ibid., p. 78.]
The canon of parsimony is a strong empirical check on scientific statement. It insists that the scientist must not affirm more than the data will allow. It thus acts as a restraint on the excitement of discovery, and it also insists that neither speculation nor unverified hypothesis is knowledge.
. . . the empirical investigator cannot be said to know what is not verified and he cannot be said to be able to know the unverifiable. Because then, verification is essential to his method, the canon of parsimony in its most elementary form excludes from scientific affirmation all statements that are unverified and, still more so, all that are unverifiable.
Ibid., p. 79. See pp. 79-83 for Lonergan’s analysis of how both classical and statistical laws satisfy the canon of parsimony.
The goal of scientific inquiry is stated by the canon of complete explanation: the empirical scientist is to strive to explain all the data. Nothing that falls within the field of inquiry (governed by the canon of selection) can be ignored, overlooked, or discarded.
Ibid., p. 84. See pp. 84-86 where Lonergan uses this canon to criticize Galileo’s theory of primary and secondary qualities and his repudiation of secondary qualities as mere appearance.
Finally, there is the canon of statistical residues. This canon arises because after classical inquiry has done its work there remain “residues” which classical inquiry cannot investigate. The existence of these “residues” calls for statistical inquiry. In short, one might argue that the canon of statistical residues arises from the imperative of the canon of complete explanation. Lonergan summarizes his general argument on behalf of the necessity of statistical inquiry in the following passage.
There does not exist a single ordered sequence that embraces the totality of particular cases through which abstract system might be applied to the concrete universe. In other words, though all events are linked to one another by law, still the laws reveal only the abstract component in concrete relations; the further concrete component, though mastered by insight into particular cases, is involved in the empirical residue from which systematizing intelligence abstracts; it does not admit general treatment along classical lines; it is a residue, left over after classical method has been applied, and it calls for the implementation of statistical method.
Ibid., p. 87. For the prior analysis that leads to this conclusion, see pp. 86-87; and for the detailed argument of this interpretation, see pp. 87-102.
The relation between classical and statistical method can be clarified by introducing the notion of abstraction, which I briefly discussed above. Lonergan’s treatment not only is clear, but summarizes and restates several key points made earlier.
So far from being a mere impoverishment of the data of sense, abstraction in all its essential moments is enriching. Its first moment is an enriching anticipation of an intelligibility to be added to sensible presentations; there is something to be known by insight. Its second moment is the erection of heuristic structures and the attainment of insight to reveal in the data what is variously named as the significant, the relevant, the important, the essential, the idea, the form. Its third moment is the formulation of the intelligibility that insight has revealed. Only in this third moment does there appear the negative aspect of abstraction, namely, the omission of the insignificant, the irrelevant, the negligible, the incidental, the merely empirical residue. Moreover, this omission is neither absolute nor definitive. For the empirical residue possesses the universal property of being what intelligence abstracts from. Such a universal property provides the basis for a second set of heuristic procedures that take their stand on the simple premise that the non-systematic cannot be systematized. [Ibid., pp. 88-89.]
And so we see that-the existence of the empirical residue and the possibility of an inverse insight into the empirical residue are the ground for the possibility of statistical method, and the canon of complete explanation impels science toward the implementation of statistical method. At the conclusion of his detailed argument, Lonergan states that
the canon of statistical residues involves three elements, and all three can be stated only in cognitional terms. The first element is the indeterminacy of the abstract: classical laws can be applied to concrete situations only by adding further determinations derived from the situations. The second element is the nonsystematic character of the further determinations. . . . The third element, finally, is the inverse insight: if the intelligibility of abstract system is not to be had [in the empirical residue], still generality is not to be renounced; for there is the generality of the ideal frequency of events; and from such an ideal frequency the non-systematic cannot diverge in any systematic fashion. [Ibid., pp. 100, 101.]
Such, then, are the six cannons of empirical scientific method. Although Lonergan does not explicitly discuss the canons in this way, it seems that these canons taken together might be understood as a more specific statement of the dynamic structure of empirical scientific method itself. The canons, as a unity structuring the conduct of scientific inquiry, attempt to provide the disciplined environment in which insights may occur and, when they have occurred, be expressed, tested and developed to their fullest. The canon of selection defines the field of inquiry and tells the scientist how to observe it; the canon of operations tells the scientist how to go about understanding that area of inquiry; the canon of relevance defines the type of understanding being sought; the canon of parsimony demands an empirical check on the understanding, demands that every statement expressing a scientific understanding be tested; the canon of complete explanation points the scientist toward renewed observation and higher viewpoints; and finally, the canon of statistical residues might be understood as an outgrowth of the canon of complete explanation. These canons constitute the specific norms of empirical scientific inquiry, and one can understand them to be a specification for the empirical sciences of the normative nature of questioning itself.
The Complementarity of Classical and Statistical Investigations: Chapter IV of Insight
Lonergan next considers the question of “whether classical and statistical inquiries are isolated or related procedures, whether they lead to isolated or related results.” [Ibid., p. 104.] Lonergan has previously established that classical and statistical methods (heuristic structures) are parallel in structure. [Ibid., pp. 58-59, 63-66; and Thesis, pp. 69-72.] In this chapter he attempts to show that they are complementary as well. This complementarity has two aspects: Lonergan first argues that classical and statistical investigations “are complementary as types of knowing;” and then he argues that in addition to this, “there is a complementarity in the to-be-known.” [Inight, p. 104.] With regard to the complementarity as types of knowing, Lonergan’s own summary of his argument will suffice for my purposes. He argues that a complementarity of classical and statistical investigations exists
at each of the stages or components of the process of inquiry. There is the classical heuristic anticipation of the systematic; there is the complementary statistical heuristic anticipation of the non-systematic.
Next, to determine either a classical or a statistical law is to prepare the way for the determination of further laws of either type; for both classical and statistical laws pertain to a single complementary field, and to know either is to effect a mental separation between types of data that have been accounted for and types that still remain to be explained.
Thirdly, there is a complementarity of formulations; the experiential and pure conjugates of classical laws can be verified only in events; the events occur only if other things are equal; and the failure to specify the other things amounts to an unconscious acknowledgement of the non-systematic aggregate of patterns of diverging series of conditions. Inversely, as conjugates are verified only in events, so events are defined only by conjugates, and statistical laws of events can possess scientific significance only in the measure that they employ definitions generated by classical procedures.
Fourthly, there is a complementarity in modes of abstraction; classical laws regard the systematic in abstraction from the non-systematic, the relation of things to one another in abstraction from their relations to our senses; but statistical laws consider the systematic as setting bounds to the non-systematic and they are confined to the observable events that include a relation to our senses.
Fifthly, the two types of law are complementary in their verification: exact and complete knowledge of classical laws cannot successfully invade the field of statistical laws; and statistical investigations are confronted with regular occurrences that admit explanations of the classical type.
Finally, there is complementarity in the aspects of data explained by the different types of laws; data as similar are explained on classical lines; but their numbers and their distributions become intelligible only by some synthesis of statistical considerations. [Ibid., p. 114-115; see pp. 105-114 for the detailed argument.]
But beyond a complementarity as ways of knowing, Lonergan also attempts to show that classical and statistical investigations are complementary because they attend to different aspects of one reality. The affirmation of both classical and statistical investigations necessarily involves a world view. This is because “whether one likes it or not, heuristic structures and canons of method constitute an a priori. They settle in advance the general determinations, not merely of the activities of knowing, but also of the content to be known. [Ibid., pp. 104-105] Or, put another way, Lonergan argues that there must be a correspondence between knowing and the known, “and, as the known is reached only through knowing, structural features of the one are bound to be reflected in the other.” [Ibid., p. 115.] Thus the dynamic structure of empirical scientific method as Lonergan has analyzed it necessarily involves a “world view” which in its general characteristics would illustrate why classical and statistical investigations are in fact complementary. Lonergan concludes Chapter IV of Insight by discussing the world view implied by his analysis of empirical scientific method, a world view he names “emergent probability,” and in contrasting that view with several other world views from the past. [Ibid., pp. 115-139.] For reasons I will discuss at the appropriate time, I wish to defer consideration of this world view to the section of this chapter on the relation between the method of science and philosophy in Lonergan’s thought. [See Thesis, pp. 108-110.]
Lonergan continues his analysis of empirical science in Chapter V, “Space and Time,” but what is added there is of no great import for my study. I will now attend to one element of empirical scientific method which has not yet received much discussion. We have seen in some detail Lonergan’s analysis of the first three moments of scientific method: observation-description; insight; and the formulation of hypotheses. We have also seen in some detail his analysis of the necessity of testing the understanding expressed in hypotheses. But while it has been mentioned, I have not yet discussed in any detail the culminating activity of scientific testing, namely, judgment. Nor have I discussed how the understanding formulated in hypotheses becomes transformed into knowledge. To these issues of judgment and knowledge I now turn.
Understanding, Knowledge, and Scientific Method
We have seen earlier [Ibid., p. 59.] that the achievement of understanding does not terminate the dynamic force and impulse of questioning; even when a scientist has understood and has formulated that understanding in a hypothesis, the further question arises, Is it so? Is it true? Or, one might phrase the question in this way: Is that the way it really is? Is that the real? And so before discussing how it is that one can make a judgment in answer to such questions, one must first consider what the notion of the real is, since that is what one is attempting to reach in making a judgment. This Lonergan does in the important Chapter VIII of Insight, “Things.”
Lonergan draws a distinction between a “body” and a “thing,” and notes that there are two different kinds of knowing that correspond to these notions. [Insight, pp. 245-254.] Let us first consider the distinction between the notions of a thing and a body.
. . . the notion of a thing is grounded in an insight that grasps, not relations between data, but a unity, identity, whole in data; and this unity is grasped, not by considering data from any abstractive viewpoint, but by taking them in their concrete individuality and in the totality of their aspects.
Ibid., p. 246. Lonergan goes on to discuss the characteristics of a “thing”: it is conceived as extended in space, permanent over time, yet subject to change.
The notion of the thing is necessary for the continuity and development of scientific thought, Lonergan notes. Fundamentally, “scientific development involves a succession of explanatory systems,” but these systems “have to be discovered in data and verified in data.” But not just any data will do.
Accordingly, scientific thought needs, not only explanatory systems, but also descriptions that determine the data which explanations must satisfy. Moreover, scientific thought needs the notion of the thing, which has as its properties both experiential and explanatory conjugates, which remains identical whether it is described or explained, which by its identity demands a coherent explanation or set of explanations that is verifiable in the easily ascertainable data of the thing as described.
Thus, the thing is the basic synthetic construct of scientific thought and development. [Ibid., pp. 247-248.]
In short, the thing is the intelligible unity grasped in data as individual and particular and concrete.
This notion of the thing will become clearer if we contrast it with the notion of what Lonergan calls a “body.” A body is the sensed unity of an object of animal extroversion.
Ibid., pp. 250-251. I might note in passing that in Whitehead’s vocabulary, what Lonergan calls a “body” has to do with sensation and presentational immediacy.
It is what Lonergan characterises as an “already out there now real.”
Lonergan’s technical definition of an “already out there now real” is as follows: “‘Already’ refers to the orientation and dynamic anticipation of biological consciousness; such consciousness does not create but finds its environment; it finds it as already constituted, already offering opportunities, already issuing challenges. ‘Out’ refers to the extroversion of a consciousness that is aware, not of its own ground, but of objects distinct from itself. ‘There’ and ‘now’ indicate the spatial and temporal determinations of extroverted consciousness. ‘Real,’ finally, is a subdivision within the field of the ‘already out there now’: part of that is mere appearance; but part is real; and its reality consists in its relevance to biological success or failure, pleasure or pain.” Ibid., p. 251.
It is the perceived unity and reality of all the objects of our common, everyday lives: the chairs we sit in, the dishes and utensils with which we eat, the walls we avoid walking into, the automobiles we drive, and so on. A body is the rock kicked by Dr. Johnson, the well into which Thales fell, the first of Eddington’s famous “two tables.”
See Arthur Eddington, The Nature of the Physical World (Cambridge: Cambridge University Press, 1928 ,”Introduction.” The distinction Eddington draws in these Gifford lectures between his “ordinary” and “scientific” tables corresponds exactly to Lonergan’s distinction between a “body” and a “thing.”
The point of Lonergan’s distinction between a thing and a body is not to deny reality to either. In fact, as Lonergan would grant, “biological success or failure” depends on the ability to perceive and deal with the objects he names “bodies,” a truth that common sense has no trouble grasping.
See Insight, p. 251. Common sense, as Lonergan describes it in Chapters VI and VII of Insight, is at home with the notion of a “body,” but would find the notion of a “thing” confusing. Eddington (see previous note) made the same observation.
Both things and bodies are real, but their reality has quite different criteria. The reality of bodies and the “knowing” of them does not depend in any way on specifically human intelligence; humans share with all other animals the perception of bodies, the biological awareness of the reality of the world of our environment. But only humans know the reality of things, and they know things only when they are exercising their intelligence. This is really the heart of Lonergan’s point. The real for the human knower is contained in the canon of parsimony:
. . . the real is the verified; it is what is to be known by the knowing constituted by experience and inquiry, insight and hypothesis, reflection and verification. Our present point is that, besides knowing in that rather complex sense, there is also “knowing” in the elementary sense in which kittens know the “reality” of milk. [Insight, p. 252.]
The differences between the two kinds of knowing should now be fairly obvious. The knowing related to bodies has nothing to do with questioning. Kittens do not question the reality of their food, nor do we. But when humans who are scientists are knowing as scientists, questioning is at the heart of the method of knowing [Ibid.] and, while human intelligence is being exercised, the real is what is to be known by verification in the data, not a perceived “already out there now.”
Lonergan proceeds in the remainder of Chapter VIII to develop his position. Since I will have occasion in Chapter III to consider those developments, I omit mention of them here. An interesting point I might note, however, is that on the basis of his position on knowing and the real, and of his notion of the “thing,” Lonergan separates his position from those of Galileo, Newton, and Kant (see ibid., pp. 252-254). This is quite similar to the way Whitehead, using his analysis of abstraction and the fallacy of misplaced concreteness, separates his position from Newton’s and Kant’s. I will return to this point in a later section.
Still the question remains, how does one come to knowledge of the real; how does one verify an understanding in the data? Lonergan pursues the answer to such questions in Chapters IX and X of Insight in which he discusses the notion and elements of judgment. He begins by relating the notion of judgment to questions. Questions, he notes, fall into two main classes: questions for intelligence (what is it?), and questions for reflection (is it?). Questions for intelligence demand answers that are either descriptions or explanations; but questions for reflection demand answers that are either affirmations or negations (or admissions of ignorance).
Insight, pp. 271-272. Lonergan also notes that a judgment involves a personal commitment and entails a personal responsibility for the judgment on the part of the one judging; ibid., p. 272. This important point is not developed until later in Lonergan’s analysis.
Next, Lonergan relates the notion of judgment to the general structure of cognitional process as he has analyzed it. [Ibid., pp. 272-274.] His prior analysis has discovered three levels in that process: the level of presentation (data); the level of intelligence (insights and formulations); and the level of reflection (judgment). In empirical scientific method these are the moments of observation and description, insight and hypothesis formation, and testing and verification. Each level leads to the next by the dynamic of questioning. Also, the second level presupposes and complements of first, and the third level presupposes and complements the second (and so the first as well). Thus the cognitional process is cumulative in character. The levels can be schematized in the following manner. [Ibid., p. 247.]
I. Data; Perceptual Images Free Images Utterances
II. Questions for Intelligence Insights Formulations
III. Questions for Reflection Reflection Judgment
Having related judgment to the three levels of cognitional process, having located it as the terminus of the process of knowing, Lonergan goes on in Chapter X of Insight to discuss specifically how one makes a judgment, that is, he discusses the operations involved in reflection. But before taking up this discussion, two points must be raised. The first is simply to note that Lonergan also distinguishes two modes of cognitional process: the direct and the introspective modes. The direct mode deals with data of sense, and follows the by now familiar dynamic of questions leading to insights, which are expressed in formulations, which lead to questions for reflection, and culminate in judgments. Empirical science, Lonergan notes, follows the direct mode of cognitional process. On the other hand, there are also data of consciousness, and the introspective mode deals with such data, following the same structural dynamic pattern of cognitional process. [See ibid.] Both these modes unfold in the three levels of consciousness. This distinction between the two modes of cognitional process is quite important in the development of Lonergan’s position, but this is not the place to discuss it. I defer such discussion initially to the later section of this chapter on the relation between scientific method and philosophy in Lonergan’s thought, and will return to it at more length in Chapter III.
A second point must be discussed to clarify a possible source of confusion. The three levels of cognitional process as Lonergan analyzes them are to be distinguished but not, in my judgment, sharply separated. Level I (data and perceptual images; free images; utterances) is on occasion described by commentators as if it had nothing to do with intelligence at all, as if it were very sharply separated from the inquiring intelligence at work in Level II.
I have in mind specifically the analysis of Kuester, “The Epistemology of Michael Polanyi,” 2: 557-560.
One commentator has even equated Level I with the kind of knowing that takes place when one “knows” a “body.”
Ibid., 2: 554-555; see also 547-549. I believe this misunderstanding is the cause for the difficulties Kuester poses, ibid., 2: 557-560.
Now it seems clear for several reasons that Level I is not to be equated with the kind of knowing that takes place when one knows a body. First, Lonergan is here discussing human cognitional process, and when discussing the “knowing” of bodies Lonergan clearly states that such knowing has nothing to do with the questioning which is the sine qua non of truly human knowing and which he describes in his analysis of cognitional structure. The knowing of a body
is constituted completely on the level of experience; neither questions for intelligence nor questions for reflection have any part in its genesis; . . . On the other hand, in fully human knowing experience supplies no more than materials for questions; questions are essential to its genesis . . . [Insight, p. 252.]
In short, the “knowing” that takes place when one “knows” a “body” cannot be called cognitive knowing, and Lonergan is here describing the levels of cognitive knowing. Further, if one applies this structure of levels to empirical scientific method, it seems clear that intelligence must play some role even on Level I because observation and description are to be located on this level (corresponding to “free images” and “utterances” respectively) and scientific observation and description are clearly intelligent activities. That is” they are not the necessary outcome of mere experience but are precise and technical operations guided by human intelligence, technical responses to the question “what is it?” Finally, while Lonergan does not discuss the presence of active intelligence on Level I at any length, he does indicate its presence in a single sentence: “The exception lies in free images and utterances which commonly are under the influence of the higher levels before they provide a basis for inquiry and reflection.”
Ibid., p. 274. The sentence occurs immediately after Lonergan presents the schematization of the three levels of cognitional process and states that each higher level presupposes and complements the levels) below it. By calling “free images” and “utterances” “the exception,” Lonergan clearly means that prior intelligent activity (Levels II and III) is often required before the operations of Level I can be conducted properly. This would be true in the empirical sciences where observation and description are often under the influence of hypothesis, and certainly presuppose past intelligent experience in scientific inquiry. Also see the quotation on p. 69 note 3, where Lonergan seems to presume some such understanding of the influence of Levels II and III on observation.
In short, in human knowing intelligence is at work in all three levels of cognitional process. While the three levels are to be distinguished clearly so as to identify and come to some understanding of the distinct operations and their patterned relationships in the dynamic structure of cognitional process, these distinctions are not to be interpreted as constituting a gulf between the three levels, separating them sharply from each other. The process of human cognition is a continuum; the distinctions are made to enable understanding of that continuum.
This interpretation, I submit, would make Kuester’s disagreement with Lonergan and his reinterpretation (see “The Epistemology of Michael Polanyi,” 2: 557-560) unnecessary, since Kuester argues that knowing is better conceived as a continuum. The issue that lies at the root of Kuester’s discussion is the relation of knowing to experience in Lonergan’s thought. I will discuss this issue in Chapter III.
With this understanding of the three levels of cognitional process, I now turn to the question of what is involved in the making of a judgment. From the schematization Lonergan presents, note that between questions for reflection and judgment occurs the activity or set of operations Lonergan calls reflection. It is this set of operations that provides the final element in the understanding of how the human knower arrives at knowledge, how the understanding which is the product of Level II is transformed into what can properly be called knowledge.
“The act of reflective understanding [reflection] is an insight,” Lonergan says, which “grasps the sufficiency of evidence for a prospective judgment. The problem is to understand what constitutes that “sufficiency of evidence.” There is “a marshalling and weighing of evidence” that takes place between the question for reflection and the judgment, “but what are the scales on which evidence is weighed? What weight must evidence have, if one is to pronounce a ‘Yes’ or a ‘No’?” [Ibid.]
To answer this question Lonergan begins by considering the general form of reflective insight. “To grasp evidence as sufficient for a prospective judgment is to grasp the prospective judgment as virtually unconditioned.” [Ibid., p. 280.] Lonergan distinguishes between the formally unconditioned and the virtually unconditioned.
The formally unconditioned has no conditions whatever. The virtually unconditioned has conditions indeed but they are fulfilled.
Accordingly, a virtually unconditioned involves three elements, namely:
(1) a conditioned,
(2) a link between the conditioned and its conditions, and
(3) the fulfilment of the conditions.
Hence a prospective judgment will be virtually unconditioned if
(1) it is the conditioned,
(2) its conditions are known, and
(3) the conditions are fulfilled.
By the mere fact that a question for reflection has been put, the prospective judgment is a conditioned; it stands in need of evidence sufficient for reasonable pronouncement. The function of reflective understanding is to meet the question for reflection by transforming the prospective judgment from the status of a conditioned to the status of a virtually unconditioned; and reflective understanding effects this transformation by grasping the conditions of the conditioned and their fulfilment. [Ibid.]
This is the general form of a virtually unconditioned judgment.
The pivotal role of reflective understanding is to grasp the conditions of the prospective judgment (i.e., “it will be so if such and such is the case”), and to grasp also that those conditions are fulfilled (i.e., “such and such is the case”). When such an insight has taken place, then the judgment is virtually unconditioned and can be made (i.e., “it is so”). It is to be noted that this grasping of the conditioned and its fulfilled conditions is an insight; that is, it is not a judgment. It is expressed in a judgment that, indeed, the conditions are fulfilled, and this results in another judgment answering yes or no to the original question for reflection. But the actual grasping of the conditioned and the fulfilled conditions is an act within the cognitional structure that has the characteristics of an insight. It is also important to note that the fulfilling conditions are another set of data; that is, in addition to the data that originally gave rise to the inquiry, the fulfilling conditions are to be found on Level I of the cognitional process (the level of presentations). There is an insight, then, into both sets of data connecting them to the same conjugates (or formulated understanding, or hypothesis). Then the further insight which is reflective understanding grasps the whole (both sets of data and the insight connecting them to a formulated understanding)—it grasps the whole as a virtually unconditioned grounding a judgment.
For the analysis behind this entire paragraph, see ibid., pp. 281-283.
Still, how does one know that such insights, which act as the pivot between questions for reflection and judgments, are correct? Lonergan resolves this question by distinguishing between “vulnerable” and “invulnerable” insights. “Insights are vulnerable when there are further questions to be asked on the same issue.” These further questions can lead to further insights that may modify, complement, or cause revision of the initial insight and explanation. “But when there are no further questions, the insight is invulnerable.” [Ibid., p. 284.] It is invulnerable because only through the development of further questions on the same issue can any modification of the insight occur. Lonergan argues that this analysis reveals “a law immanent and operative in cognitional process.” His analysis reveals that prior to a conceptual distinction between correct and incorrect insights there can be found an “operational distinction” between vulnerable and invulnerable insights. The “immanent law of cognitional process” that may be formulated from this analysis is that “such an insight [grasping the fulfilment of conditions of the conditioned] is correct, if there are no further, pertinent questions. [Ibid.] How does one judge that there are no further pertinent or relevant questions? Lonergan here appeals to what he calls the “self-correcting process of learning”:
Judgment on the correctness of insights supposes the prior acquisition of a large number of correct insights. But the prior insights are not correct because we judge them to be correct. They occur within a self-correcting process in which the shortcomings of each insight provoke further questions to yield complementary insights. Moreover, this self-correcting process tends to a limit. . . . that self-correcting process reaches its limit in familiarity with the concrete situation and in an easy mastery of it. [Ibid., pp. 286-287.]
Lonergan goes on to exhibit how this analysis resolves the problem of induction [Ibid., pp. 287-289; see also Thesis, p. 118.] and how it reveals common sense and empirical science to be separate universes of discourse using essentially the same process but operating with different standards and criteria. [Ibid., pp. 289-299.] I will take up these issues at later points in my study.
Let me summarize, then, how this analysis applies to empirical scientific method. Earlier I discussed the elements of empirical scientific method that correspond to Levels I and II of the cognitional process: Level I is the marking out of the field of inquiry, observing the data, and describing them. Out of this initial moment arise questions for understanding, or what Lonergan also calls questions for intelligence. These are the specific problems to be solved, and these constitute the initial moment of Level II of the cognitional process. Insights occur, which are the private understandings of the solutions to the problems, and these insights are then expressed and formulated in hypotheses. The hypotheses are the public statements of the understandings grasped in the insights, and such expressions constitute the third and final moment of Level II. But the scientist is not satisfied with explanatory hypotheses. He or she is driven, personally and by the canon of parsimony, to ask “is it so?” The asking of such questions for reflection constitutes the initial moment of Level III of cognitional process. In this subsection I have examined what generic operations are involved in the terminal activities of empirical scientific method: deduction, forecast or prediction, the devising of experiments, and the testing of the hypothesis by the performance of the experiments. I have also tried to identify what cognitive operations are involved when a scientist declares that the performed experiments verify (or falsify) the hypothesis being tested. In terms of Lonergan’s analysis, the operations of deducing the implications of the hypothesis and predicting what will happen in certain circumstances according to the hypothesis are operations geared toward discovering the conditions of the prospective judgment. Devising experiments to test the hypothesis is determining how one can arrive at the data that will exhibit whether or not the conditions are fulfilled. Finally, the performance of the experiments provides the data in which those conditions are in fact fulfilled or not fulfilled. These are all operations of reflective intelligence. When the results of the experiments are available to the scientist there occurs in him or her the insight of reflective understanding grasping the conditions of the prospective judgment and their fulfilment (or their lack of fulfilment). This insight is then expressed in the statement (judgment) that the hypothesis has been verified (or falsified).
Verification, then, is in the end a judgment. The making of this judgment depends on the performance of a prior appropriate pattern of acts or operations commonly referred to as checking, testing, or verifying. [See ibid., pp. 326-327.] That appropriate pattern of operations determines the conditions of the judgment, and the specific way of discovering if the conditions are fulfilled. It generates a new set of data, and requires an insight linking both new and originating data to the hypothesis. This insight results in a judgment that indeed the conditions are fulfilled or that they are not. This insight and judgment in turn enable the further insight which is the reflective grasp of the virtually unconditioned. This latter insight grounds the judgment that, yes, the hypothesis is verified or, no, the hypothesis is not verified. If there are no further pertinent or relevant questions, what was understanding (at the final moment of Level II) has been transformed by judgment into knowledge (at the final moment of Level III). If there are no further pertinent questions, one now not only understands but also knows. One has reached the real by verifying the understanding in the data, by affirming it in an act of judgment.
This analysis reveals why empirical scientific method can be called a “normative pattern of recurrent and related operations.” This pattern is normative because, as we have seen, the making of a judgment (verifying) depends on the performance of the whole pattern of operations: the judgment depends not only on the patterned operations of reflective understanding (Level III), but also on the patterned operations of Levels II and I. The judgment simply cannot be made responsibly unless all the other acts or operations occur and recur in precisely this pattern. This normative pattern, it is to be noted, is essentially the expression of the nature of inquiry itself. That is, the norm is not external and applied to the cognitional process, but rather is carried internally in the dynamic of questioning. The demands of this questioning can only be met by asking and answering the questions in this way, and the demands with which questioning confronts the inquirer only cease when one can finally make a virtually unconditioned judgment.
This analysis also reveals why empirical scientific method yields “cumulative and progressive results.” I said above that judgment transforms understanding into knowledge if there are no further pertinent questions. But as anyone familiar with scientific work knows, there are always further pertinent questions (whether the scientist is immediately aware of them or not). This fact at once accounts for the cumulative character of scientific knowledge and also for its limited or partial character, and to these topics I now turn.
The Nature of Scientific Knowledge
Throughout his writing, whenever Lonergan discusses modern science he always comments on the limited character of scientific knowledge in contrast with the classic Aristotelean ideal of science
In addition to the reference in the following note, see these articles, all in Second Collection: “The Future of Thomism, p. 51; “Belief: Today’s Issue,” p. 94; “The Absence of God in Modern Culture,” pp. 103-104; “Theology and Man’s Future,” pp. 139-140; and “Revolution in Catholic Theology,” pp. 234-235. See also Method, p. 15; and Bernard J. F. Lonergan, Philosophy of God, and Theology (Philadelphia: The Westminster Press, 1973 , pp. 6-7 (hereafter cited as PGT).
The following quotation will give some sense of Lonergan’s interpretation.
The Greek conception . . . envisaged science as true, certain knowledge of causal necessity. But modern science is not true; it is only on the way towards truth. It is not certain; for its positive affirmations it claims no more than probability. It is not knowledge but hypothesis, theory, system, the best available scientific opinion of the day. Its object is not necessity but verified possibility. . .1
Bernard Lonergan, “Dimensions of Meaning,” in F. E. Crowe, ed., Collection: Papers by Bernard Lonergan (New York: Herder and Herder, 1967), p. 259; see also pp. 260-261. (This work hereafter cited as Collection.)
In our present context the important point is that the judgments of empirical science are not absolute or certain judgments, but only probable judgments, and this is because of the constant presence of further pertinent questions. Thus the knowledge gained by a probable judgment of verification in the empirical sciences is knowledge, but incomplete knowledge; it is not an absolutely certain grasp of the truth, but only a closer approximation to the truth.
Lonergan establishes this interpretation in his analysis of probable judgments and the application of that analysis to the empirical sciences. [Insight, pp. 299-304.] He notes that probable judgments converge upon true judgments or approach them as a limit. This seems to be a paradox, for how can the probable be known to approach the certain, when the certain is unknown? [Insight, p. 300.] The answer is that the knower is able to recognize when the truth is being approached, and what enables that recognition is the “self-correcting process of learning.”
As we have seen, the self-correcting process of learning consists in a sequence of questions, insights, further questions, and further insights that moves towards a limit in which no further pertinent questions arise. When we are well beyond that limit, judgments are obviously certain. When we are well short of that limit, judgments are at best probable. . . . In brief, because the self-correcting process of learning is an approach to a limit of no further, pertinent questions, there are probable judgments that are probably true in the sense that they are approximate to a truth that as yet is not known. [Ibid. See also Thesis, pp. 88-90.]
Applying this analysis to the empirical sciences results in the conclusion that scientific judgments are no more than probable. They can be no more than probable because an essential part of making a scientific judgment (verification) is a return to the concrete in testing, and the concrete always raises or may raise further’ pertinent questions.
Empirical science gets its start by hitting off significant correlations. The correlations implicitly define abstract correlatives. But precisely because they are abstract, the return to the concrete is greeted with further questions. . . .
The generalization of classical laws, then, is no more than probable because the application of single laws raises further questions that head towards the systematization of a whole field. In turn, such systematization is no more than probable until the limit of no further, pertinent questions is reached. But that limit is not reached, first, if there may be further unknown facts that would raise further questions to force a revision or, secondly, if there may be further, known facts whose capacity to raise such further questions is not grasped.
Similar considerations render the generalization of statistical laws no more than probable.
Insight, pp. 301, 302. Lonergan there gives examples from empirical science for further clarification.
Hence a scientific judgment always carries with it a qualification: this hypothesis is verified, so far as we know at present or so far as we can tell at present. The scientific judgment does result in knowledge, but it is a knowledge that is limited, partial, and always open to revision in the future.
Yet the same presence or possibility of further, pertinent questions which reveals the limited character of scientific knowledge, also reveals the cumulative and progressive character of scientific knowledge.
Questions yield insights that are expressed in hypotheses; the testing of hypotheses raises further questions that generate complementary insights and more satisfactory hypotheses. For a while the process advances in widening circles; then the coherence of system begins to close in; investigation turns from fresh ventures in new fields to the labour of consolidation, of working out implications fully, of settling issues that leave the general view unchanged. The self-correcting process of learning is palpably approaching a limit. [Ibid., p. 303.]
But even in such situations in a developing science, new questions can arise, new ways of considering old data, new questions that provoke the perception of new data or that cause the scientist to see new significance in old data. The questions raised in testing old hypotheses may generate a new set of insights, resulting in a basic revision in the science; the old hypotheses, laws, and standards are found to be insufficient and new hypotheses, laws, and standards are worked out. [See Ibid., p. 166.] Clearly progress in understanding has been made, and one can characterize that progress as cumulative because the new insights, hypotheses, and laws preserve “all that was valid in the old.” [See Method, p. 5.] For example, Einstein’s theory of relativity does not totally discard Newton’s theory of gravitation but rather, even while exhibiting its limitations, includes it within the new theory of gravitation as a special case. There has been progress in understanding and in knowledge, and that progress is cumulative. The scientist can say that while we have not yet grasped the truth completely, at least we are closer to it now than we were before. Thus the cumulative and progressive advance of scientific knowledge defines an asymptotic approach to the truth.
In Insight, pp. 303-304, Lonergan suggests that there are lower and upper limits to the progress of science. The lower limit is defined by the ability to perceive sensible differences in the data. The upper limit is defined by the invariant structures of cognitional process which, Lonergan argues, imply a limit to the variation of theoretical constructions. Since he takes up this topic in his analysis of metaphysics, I will defer comment on this issue until Chapter III.
With this we have seen the last of the main points in Lonergan’s interpretation of empirical scientific method.
A summary of Lonergan’s interpretation of empirical scientific method does not seem necessary here since I began my study with Lonergan’s own summary statement. See Thesis, pp. 56-57 or Method, pp. 4-6.
It has been necessary to treat his interpretation at length in order to exhibit the thoroughness with which Lonergan considers the elements of scientific method and also because in his analysis of it Lonergan establishes several conclusions of great significance for his interpretation of philosophy and, ultimately, of method in theology. It is now time to attend to the relation between empirical scientific method and philosophy in Lonergan’s thought.
Forward to Lonergan's Interpretation of Scientific and Philosophic Method: The Method of Empirical Science and Philosophy