HOW WE KNOW WHAT WE THINK WE KNOW

To a large extent, life consists of overcoming the problems we encounter in our attempts to achieve our purposes. Along with the easy problems in life are many enormously complex and difficult ones. These would be considerably less difficult if our notions about how the world works were more reliable.

It is comforting to have reliable knowledge to deal with problem situations that have straightforward, linear cause-and-effect relationships. For example, fixing a flashlight that no longer works by replacing the batteries poses little challenge to our knowledge of cause and effect. But, approaching complex problems with an overly simplistic linear mindset often makes matters worse instead of better.

Based on an analysis of the work of people, especially scientists, who have been extremely successful in solving complex problems, I have learned three lessons that are important to a better understanding of knowing:

1. Reality as we know it is just our perception of it—a kind of map of reality, not the true territory of reality.

2. Action is an integral part of cause-and-effect loops, with purpose playing a critical and often-overlooked role.

3. Identifying the strongly held assumptions (beliefs) that influence what we perceive and how we determine our actions in the world is vitally important to opening us up to perceiving new feedback information and to faster knowledge improvement.

Putting these lessons into practice takes conscious effort, because much of our life experience has been dealing with the outside world as independent components of reality for which one-way, or linear, cause-and-effect thinking is adequate.

THE PAK (PERCEIVING-ACTING-KNOWING) LOOP

The perceiving-acting-knowing system can be visualized as a loop of intimately related components. Figure 1.1 illustrates the components of this system, which I refer to as the PAK Loop. A useful understanding of how this system functions requires a focus on the loop as a whole and not on the components in isolation.

As noted by the psychologist Hadley Cantril in the quotation at the beginning of this chapter, perceiving, acting, and knowing is an interdependent process. Nevertheless, a discussion of the PAK Loop requires some starting point. For convenience, we will begin at the point where an individual is trying to achieve a purpose within the context of the perceived world “out there.”

Purposes

Purposes are personal. They are the outcomes we, as individuals, seek from the actions we take. (This is not to say we always get what we seek.) The great bulk of our purposes are mundane. Consider all the specific, detailed purposes and related actions taken in driving to work—from as small, or low-level an action as moving the steering wheel a little to the left or right to counteract a crosswind so the car stays on our intended course. Some larger, or higher-level, purposes of driving to work would include: why you work (survival? self-fulfillment? enjoyment?) and why you have a particular job (steppingstone to a better job? prestige? power?). It quickly becomes evident that we function within a hierarchy of purposes, with higher purposes guiding, or setting, lower purposes.

FIGURE 1.1 PAK Loop

Source: Madden (2008b).

Being cognizant of higher-level purposes is especially relevant to business wealth creation. For example, in Chapter 7 I describe the decision of a Japanese pharmaceutical company’s top management to align the firm ’s mission statement (purpose) with the higher-order purpose of genuinely helping patients that was widely shared by employees. One result was significant improved corporate financial performance.

Studies of brain activity suggest that many of the common things we do are not associated with brain areas that are responsible for awareness or consciousness. Apparently, we operate much of the time as if on autopilot (Gazzaniga, Ivry, and Mangun, 2008). This is highly functional, and indeed necessary. Otherwise, our consciousness would be overwhelmed by minutiae—perceptual noise. Evolution has equipped us to do things much more quickly than we could if everything required conscious mental processing. Many actions would be impossible. Think of all the things that require virtually instantaneous “muscle memory,” such as getting out of bed, walking, or typing.

But being on autopilot has its downside. Consider two economists given the task or purpose of evaluating whether minimum wage legislation is good or bad for the economy. One economist is a believer in free markets and the other believes government regulation is necessary to prevent or fix market deficiencies. Because of their core assumptions, they are on different automatic pilot programs, and their expectations are already set to a large degree (Olson, Roese, and Zanna, 1996). The data they choose to consider (and ignore), the time periods covered, and the forms of analysis employed for the lower-level research purpose of evaluating the economic impact of minimum wage legislation are most likely to be biased.

Economists (and other inquirers) who have a genuine, higher-level purpose of better understanding cause and effect need to explicitly guard against being guided by their automatic thinking and acting templates. Such researchers would be well served by, at an early stage, explicitly working creatively to overcome the heavy hand of often-unconscious beliefs.

Perceptions

Any discussion of perceptions raises the age-old philosophical question, “What is reality?” (Madden, 1991). Thinking that there is a pure, independent reality needs to be replaced with the concept that reality is actually dependent on an individual’s past experience and current knowledge base, such that each of us is a participant in perceptions of what is “out there.” This also helps put into practice one of the hallmark criteria of science, namely, that all knowledge is tentative and subject to revision.

In the 1940s and 1950s, Adelbert Ames Jr. and his colleagues initiated a paradigm shift away from the view of perception as a passive response to the external environment and toward the view of perception as a process actively carried out by the individual (Bamberger, 2006). Ames was frequently labeled a genius due to his path-breaking research in visual perception at the Dartmouth Eye Institute. Ames and John Dewey often exchanged ideas on Dewey’s transactional approach to knowing as it related to perception (Cantril, 1960).

The Ames Demonstrations were a series of ingenious laboratory experiments that illustrated the dominating influence of observers’ strongly held assumptions. For example, assumptions that floors are level, windows rectangular, bigger is closer, and the like, are particularly strong because of our extensive experience with actions being successful based on the validity of these kinds of assumptions. When an experiment falsifies a strongly held assumption, we nevertheless construct a visual “reality” that conforms to what we “know” to be true.

The Ames Demonstrations in visual perception were instrumental in showing that purpose, perception, and action are all parts of a single connected system.¹

[T]hese experiments . . . suggest strongly that perception is never a sure thing, never an absolute revelation of “what is.” Rather, what we see is a prediction—our own personal construction designed to give us the best possible bet for carrying out our purposes in action. We make these bets on the basis of our past experience. When we have a great deal of relevant and consistent experience to relate to stimulus patterns the probability of success of our prediction (perception) as a guide to action is extremely high, and we tend to have a feeling of surety. When our experience is limited or inconsistent, the reverse holds true. . . . [P]erception is a functional affair based on action, experience and probability. The thing perceived is an inseparable part of the function of perceiving, which in turn includes all aspects of the total process of living.

(Ittelson and Kilpatrick, 1951, p. 55)

The interdependent processes that contribute to visual perception are analogous to the components of the PAK Loop, which are best viewed as cross-linked together in a system that, for the most part, operates simultaneously as opposed to a mechanistic step-by-step procedure.

Cause and Effect

Problems are perceived within a given context. Attention to context increases as one’s knowledge base broadens and one is able to appreciate ever-greater complexities of cause and effect. This leads to wider avenues for drawing on patterns that were adequate in the past for connecting cause to effect. Some patterns, or assumptions, have proven so reliable in the past that we take them as non-debatable truths. For example, when driving we use assumptions about the size of cars. Consequently, when approaching cars are seen as getting bigger, we also perceive them as getting closer.

Experts have more patterns to draw on than do non-experts. When past experience seems insufficient (as with a new problem), one looks for additional information (creating a new purpose) and that can lead to hypotheses about a root cause. How a problem is formulated, the initial selection of variables to study, the first hunch at possible connections, and the criteria used for evaluating the evolving hypotheses do not arise in an objective, unbiased fashion (Argyris and Schön, 1996).

In analyzing cause and effect, decision makers need to be keenly aware of the deep pull of their existing knowledge base about how the world works, which has been built up over a lifetime of experience. Also, decision makers should be attentive to the organization’s culture or way of doing things that has evolved to meet a variety of purposes that, in subtle ways, may interfere with the primary goal of the organization. Culture results in strongly held assumptions that influence how problems are perceived and the extent to which hypotheses about cause and effect need testing.

Consider two examples with horrific consequences due to faulty analysis of cause and effect:

1. Will the cold temperature at liftoff cause failure of the O-ring seals for the rocket that propels the Challenger space shuttle?

2. Will damage from the observed foam debris at liftoff for the Columbia space shuttle impair reentry?

The Columbia Accident Investigation Board approached their work with a systems mindset. The Board concluded for both disasters that “previous political, budgetary, and policy decisions . . . impacted the Space Shuttle Program’s structure, culture, and safety system . . . these in turn resulted in flawed decision-making for both accidents” (CAIB, 2003, p. 195).

That improved cause-and-effect analysis leads to better decision making, there is little doubt. But cause-and-effect analysis is not performed in isolation, even though one might, at times, believe otherwise. Rather, the analysis of cause and effect is best viewed as one component of the PAK Loop.

Actions and Consequences

The purpose of analyzing cause and effect is to learn to take actions that will yield desired consequences. As systems become more complex, so, too, does cause and effect.

Particularly in economic matters, decisions can have decidedly different near-term and long-term effects. A classic public policy example is when government officials employ an easy credit and money policy to stimulate near-term general income, output, and employment. Only sometime later do the negative effects appear in the form of rising prices and cyclical corrections of unsustainable resource allocations. A similar time delay of effects has been observed when a new CEO, noted for cost-cutting, makes large cuts in a firm’s R&D budget and fires talented employees in order to improve near-term accounting earnings. But the loss of employee trust and talent reduces the firm’s ability to create long-term wealth. The key point is that effects can occur with or without a time lag, or in a different physical location from the original cause, leading to erroneous conclusions about the consequences of particular actions.

Let’s return to the foam debris issue that damaged the Columbia