All Life Is Problem Solving

One of the concepts George Soros emphasizes the most is “reflexivity.” Here’s his presentation of the idea from The Age of Fallibility (pp. 6-7).

”On the one hand, we seek to understand our situation. I call this the cognitive function. On the other hand, we seek to make an impact on the world. I call this the participating function. The two functions work in opposite directions and they can interfere with each other. The cognitive function seeks to improve our understanding. The participating function seeks to improve our position in the world. If the two functions operated independently of each other, they could in theory serve their purpose perfectly well. If reality were independently given our views could correspond to reality. And if our decisions were based on knowledge, the outcomes would correspond to our expectations. But that is not the case because the two functions intersect, and where they intersect they may interfere with each other. I have given the interference a name: reflexivity. . . .”

In The Crash of 2008, Soros called the participating function, the manipulative function. A number of issues arise from this statement of the reflexivity idea. First, what does it mean to say that “the two functions work in opposite directions . . . ”? What does it mean to say that “. . . they can interfere with each other”? Soros provides many examples at other places in his work of “reflexivity.” But I believe that none of these examples makes clear what the terms “intersect” and “interfere” mean. I’ll return to this point later on.

Second, Soros says that “If reality were independently given, then our views could correspond to reality.” But, couldn’t reality be independently given, and still be beyond our understanding so that our views of it don’t correspond to reality? Aren’t there many aspects of reality that humans have no impact on, that our scientific knowledge doesn’t handle very well?

Also, this phrase assumes that reality is not independently given. Yet, surely there is a middle ground between reality being independently given and its being determined by the manipulative function. Specifically, certain aspects of reality are clearly independent of human actions. Other aspects of reality can be affected by our actions, but in many, many instances the influence is only very small, even hardly measurable. In still other aspects, we, and others, can have substantial influence on the course of reality. And, in still others, we may find ourselves in positions where our decisions can directly impact reality and we correctly predict and perhaps even control, an intended impact, though even in these instances many of the consequences of our decisions may be unexpected and unintended. So there are many degrees of “independently given” where we can do little to immediately affect reality, even when we are focusing on social reality, and not on material reality that is beyond the social.

Further, the fact that our actions can impact reality, i.e. that we can influence what will happen, either substantially or only slightly, doesn’t exclude the logical possibility in any given situation that we can assess it correctly and state the truth about aspects of it, without having or providing a complete description of it. So, the fact that we can impact our situation doesn’t imply that our assessment of it before having an impact on it won’t correspond to the way it is before we act.

Third, in saying that “if our decisions were based on knowledge, the outcomes would correspond to our expectations,” isn’t Soros assuming that truth is a requirement for using the term “knowledge?” Of course, the classical definition of knowledge is that it is “justified true belief.” So, truth is a requirement on the classical account; but Soros’s mentor, Popper, denies this JTB account of knowledge. In his view, knowledge can be false, and, in fact, most, or at least much, of the time it is. So, on Popper’s view, which I share, and which I’m surprised Soros evidently doesn’t, since it is one of the epistemological bases of Popper’s Open Society and Its Enemies, even if our decisions are based on “knowledge,” their outcomes will frequently not match our expectations. Not only that, but such mismatches are critical to learning and to the growth of our knowledge, because they give rise to problems and, for Popper, solutions to problems that we have tested, and not yet refuted, are new knowledge.

Fourth, returning to the idea of “reflexivity” as “interference” of the two functions with one another, Soros, provides many examples of what it means, and sees the idea of “interference” as going back to antiquity. But the idea cries out for more development, analysis, and specificity, to clarify the notions of “interfere,” and “intersect.” He really doesn’t deliver this in his work. So, we can’t really tell whether his quite successful forecasts of problems in the economy, are due to his conceptual framework as he explicates it, or whether that success is due to other notions he uses apart from “reflexivity.”

To get a better idea of what’s involved in the ideas of “reflexivity,” and “interference,” let’s look at some of the historical development of ideas closely related to reflexivity before Soros. The idea of reflexivity has been around for some time, and is well-recognized in the philosophy of science. This wikipedia entry provides a summary of at least part of the modern history of the idea.

The article indicates that the notion originates much before the 1970s and mentions W. I. Thomas, a noted sociologist, as the originator of the core idea in the 1920s. Robert K. Merton, named one pattern of reflexivity “the self-fulfilling prophecy”, in a famous article appearing in 1948. John Dewey discussed the problem in The Public and Its Problems, first published in 1927. And David Easton discussed both the self-fulfilling and “self-denying” prophecies in his well-known, The Political System. Ernest Nagel treated the topic well in his 1961 text, The Structure of Science. The wikipedia article emphasizes the sociological origins of the idea, and also mentions Soros. But I think its view of reflexivity may be too restrictive.

The article also mentions Karl Popper’s treatment of “the oedipal effect” in The Poverty of Historicism, 1957. However, The Poverty was first published in installments in Journal form in Economica in 1944-45, while Popper was still in New Zealand, and both preceded Merton’s famous article, and may well have been available to Merton. Also, Popper actually named his reflexive pattern “the oedipus effect.”

Popper’s account also raises the question of whether reflexivity originates with Thomas or has an older origin. In The Poverty, Popper attributes the idea of reflexivity to the historicists, and this suggests that it may have originated with the Greeks. Sure enough, Barry Sandywell has written a book on Pre-Socratic Reflexivity that appears to make the case that reflexivity began with Thales.

Whether this is right or not, I think it’s clear that reflexivity is far older than the 1970s. In fact, in my own Political Science Graduate School experience from 1960 – 65, self-fulfilling and self-denying prophecies were frequently, and prominently, mentioned as reasons why the behavioral approach to political science, with its emphasis on scientific method, could not possibly succeed.

All this is not to say that reflexivity is not a more serious practical problem impeding the growth of knowledge in the social sciences, as opposed to the physical sciences where we have more situations where our actions cannot change our situations. It is however, to suggest that reflexivity is not (a) a problem social scientists are unaware of (b) a fundamental ontological fact or logical condition that somehow makes the epistemology of the social sciences different from the epistemology of physical science, and (c) a factor that makes the social sciences necessarily less “objective” than the natural sciences.

The earlier treatments of “reflexivity” raise the question of whether there is a significant difference between these developments and Soros’s views, and whether that difference may not shed some light on the issue of “interference.” Pursuing this thought, the previous treatments of the self-denying and self-fulfilling prophecies, and “the oedipus effect” all see the interaction between what we think and what we do as sequential in character. That is, it takes some time for the cognitive function to affect the manipulative function, and for the manipulative function to affect the cognitive function. Soros, however, at least some of the time, views the relationship between the two functions as “simultaneous,” as indicated by his statement in The Crash of 2008 (p. 29) “A situation may be reflexive even if the cognitive and manipulative functions operate sequentially and not simultaneously.” This statement however, also indicates that sequential processes can also be “reflexive.” So, for Soros, there are at least two kinds of “reflexivity,” simultaneous and sequential. The first kind, “simultaneous reflexivity,” a two-way relationship between the two functions at the same point in time, isn’t consistent with earlier uses of the term, since all of these interactive relationships are sequential. The second kind, ”sequential reflexivity,” is consistent with earlier writings related to “reflexivity.” The question is: does “interference” look any different, depending on the kind of reflexivity one is talking about. I think so, and in my next blog I’ll discuss “interference” in both contexts.