The Basics of Logical Analysis 3: Concluding

I wanted to conclude the line of discussion I was following in my previous posts, with an eye toward the experience a researcher might have in beginning to define a new project, particularly those in areas where the researcher has not done a lot of previous research.  I also wanted to try to make my examples a little more detailed and academic in terms of focus. I’m still going to be working with an example from an area where I have little experience because it’s close to one of the concerns of a writer with whom I’m working.

Down the Rabbit Hole 4: Fractals

The previous post was talking about “going down the rabbit hole” for the way that a question can seem initially simple and small, but takes on detail and scope as it is examined more closely. Another parallel would be fractals, which are patterns/images derived from mathematical operations that are recursively defined in such a way that as you magnify the image, new detail continuously emerges. The Mandelbrot Set is one of the most famous of the fractals. 

Research shares something of this characteristic. It may not be infinitely recursive (though some have argued that it is), but generally, if you examine any issue closely, it will lead to more questions.  This is due to the basic nature of analysis: if we analyze things into separate parts/aspects/issues, each of those separate parts can itself be analyzed into its ow constituent part.  Jorge Luis Borges wrote an essay titled “Avatars of the Tortoise,” in which he argues that infinite regressions “corrupt” reasoning, giving examples, like how to define a word/concept, it is necessary to use other words, and each of those other words then needs to be defined, which requires other words, which then all require their own definition, and so on. I’m not sure that the pattern is infinite (there are, after all only a finite number of words, so for definition at least the regression can’t be infinite), but the multiplication of details can quickly become overwhelming. 

The Nobel Prize-winning psychologist and economist Herbert Simon, whose studied decision-making, coined the term “satisficing,” to speak of how some decisions must be made without a full logical analysis because such analyses take so long and become so detailed.

As my earlier examples of reviewing a restaurant or movie showed, it’s pretty natural to see different aspects in things: the restaurant has food and service and ambiance; the service has courtesy and competence; courtesy has all the different things that different people said and did. It may be simple to say whether you liked the restaurant, but to explain in detail all the different factors that contributed to that decision is another matter altogether.

Fractal: The Barnsley Fern
Each leaf, if expanded, will show similar structure and fine detail to the larger frond.
Image by: DSP-user / CC BY-SA (https://creativecommons.org/licenses/by-sa/3.0)
(from https://commons.wikimedia.org/wiki/File:Barnsley_fern_plotted_with_VisSim.PNG)

A More Focused Example

So far, I was giving pretty general examples, now let’s try to get more focused.

Let’s imagine a hypothetical student, studying business management.  And let’s imagine that this student has what we can call “The Fruit Theory of Management,” in which they assume that giving employees fruit improves performance. (I was going to call it “Apple Theory” but didn’t want this to be confused for a reference to the big corporation.)

The Fruit Theory

On its face, the fruit theory of management is ridiculous, but since I’m talking about a general structure of research, the precise theory in question is not so important (as will hopefully become obvious in a moment). Instead of “giving employees fruit” we could use “giving employees training in XYZ,” or, more generally, “instituting policies ABC.”  ”Giving fruit” can stand in for any possible intervention. And instead of “employees,” we could substitutes almost any group—students, parents, plumbers, etc.—and in each of these cases we could either find a suitable measure of performance, or we could replace “performance” with some other construct to measure (e.g., happiness, health, etc.).  

We can even generalize this to any basic causal pattern: “giving fruit leads to better performance” is a specific example of the general pattern “X causes Y.” Most research is concerned with causal relationships in some way or another, so although I’m going to focus on fruit theory

Studying Fruit Theory

So, we have our business management student who wants to research fruit theory. Generally speaking, a starting point for fruit theory would be to define the theory.

So the student tries to write down a definition (or speaks a definition in conversation with someone). At this point, the process of analysis inevitably has already begun: the words used can themselves be examined individually.  So, if the theory is “giving fruit leads to better performance,” there are elements that can be defined individually. 

For starters, we can ask “what is  fruit?”  In everyday conversation, we know what a fruit is and don’t need definition. But if we’re talking about developing research and examining causal relationships, we want to define things more closely and formally. (Research needs formality and detail so that others can check the research.)  For example, fruit theory might call for fresh, ripe, worm-free fruit that people would enjoy eating (a definition that is not identical with a more general understanding of fruit that includes unripe or wormy or rotten fruit). That might lead us to a whole set of questions of how to identify fruit that people would enjoy eating, which could lead to more general questions of what it means for people to enjoy eating. (Or maybe the real issue is that people enjoy receiving fruit as gifts—that would lead to a different definition of what “fruit” is.)

To study fruit theory, we also need to define what counts as “giving” and what counts as “better performance.”  As for “giving,” there is some question of the specific details of how the transfer is made and whether any conditions are placed on that transaction, including any potentially hidden costs of the transaction. But defining giving is relatively simple compared to the question of “better performance.” Measuring performance a huge array of questions: Whose performance? Are we measuring the performance of the organization as a whole? Or of individuals in it? What kind of performance? What dimensions of performance are we measuring (speed? accuracy? gross sales? net sales? etc.) and over what time periods? Are we measuring cash flow of the business over a month? Or the employee sick days taken over a year? Or are we measuring profitability over a decade? There are any number of different ways to think about the general concept of performance.

To develop research, we might also need to specify further the causal mechanism by which fruit theory works. Does giving fruit work because fruit makes people healthier, and therefore better able to work hard (as the old saying goes “an apple a day keeps the doctor away”)? Is there a physiological causality? Is that physiological causal path one that gives people more energy? Or one that improves their strength? Or one that boosts their mood?  Or maybe the causality is not physiological but psychological: giving employees gifts makes them feel appreciated and they want to work harder as a result?

Answers lead to new questions

Whenever we make a choice of where to focus attention, we can find new questions to pursue. We may start pursuing a question of business, as in fruit theory, but that question might lead into other fields of study.  If we posit a physiological cause for fruit leading to better performance of employees, then we need to study physiology. That study might lead in a variety of directions: maybe fruit theory works because fruit improves health, reducing sick-time lost—that would lead to study of immunology: how and in what ways do apples improve immune response? Or maybe fruit theory works because of some other physiological effect: strength, endurance, mood. Since different foods and substances can impact strength, endurance, and mood, maybe fruit has such effects?  If one thinks that fruit has a physiological effect on mood, one might then be led into questions of which specific biological pathways lead to mood improvement, and perhaps in studying that research, you see that other researchers have identified different kinds of mood improvement, and perhaps debate ways in which physiology affect mood.

New answers pretty much always suggest new questions.  

Preventing Over-analysis

You can take analysis too far. If you constantly analyze everything, you end up with a great mass of questions and no answers.  It can lead to getting swamped in doubt.  There is no rule for this, beyond that at some point it is necessary to pick the point at which you say “I’m satisfied with my answer to this question.”  Such statements close off one potential avenue of study to allow focus on another, and to set limits to what you need to study—limits that are necessary for the practical reason that it’s good to finish a project even if that project is imperfect.

If you say “I’m satisfied that the reason Fruit Theory works is because fruit makes people healthier,” you don’t need to pursue questions of whether and how and how much fruit promotes health, and you can go on to focus on how improved health helps a business.  Or if you say, “I’m satisfied that fruit theory works,” you can go on to study details of implementing fruit theory.  Of course, it’s good to have reasons, and good to be able to explain those reasons: if you’re satisfied that fruit theory works, it’s useful to be able to give evidence and reasoning. In academia, that evidence often comes in the form of other research literature. If you can cite five articles from reputable sources that all say “fruit theory works,” then you can go on to your research in implementation without getting embroiled in any debate about whether fruit theory works—even if the five articles you cite are not yet accepted by all members of the scientific community.

Conclusion

Analysis itself isn’t really that hard in the small-scale—we do it automatically to some extent. But it is something that grows increasingly difficult as we invest more energy into it: the more detail we add to our analysis, the more there is an opportunity to analyze further, which can lead to paralysis or to getting swamped. It is something that wants care; it wants attention to detail.