Toward a theory of Serendipity

What follows are some of my developing ideas on browsing and serendipity. I’m not really convinced of any of it, but it helps my own thinking to write it down, and hopefully get feedback …

The OED defines serendipity as “The faculty of making happy and unexpected discoveries by accident. Also, the fact or an instance of such a discovery.”

To me, this implies that instances of serendipity vary along at least 2 dimensions — How happy they make you, and how unexpected they are. If we limit the discussion to serendipity in the service of research, then we can think of “how happy they make you” as “how useful they are.” So, two essential qualities of serendipity are usefulness and unexpectedness. The holy grail of serendipity is the essential, but wholly unexpected discovery. Most of our serendipitous discoveries fall somewhere short of that on one or both dimensions.

Since social scientists love 2 x 2 tables, I think one is called for here:

Dimensions of Serendipity

	Low Usefulness	High Usefulness
Low Unexpectedness	Why Bother?	Only slightly off the beaten path but essential to your research agenda
High Unexpectedness	Nice reference to show you are thinking broadly, but not essential to your main research agenda	Holy Grail of Serendipity

Thinking about serendipity in this way helps me get past the simplistic debate over whether print browsing or online browsing is better for serendipity.
The more interesting question for me is: What conditions are likely to produce what kinds of serendipity?
If we think about the relationship between browsing and serendipity, one reasonable set of hypotheses might be:

H₁: The larger the body of materials being browsed, the higher the degree of unexpectedness.
H₂: The more ordered/organized the body of materials being browsed, the lower the degree of unexpectedness.
H₃: The larger the body of materials being browsed, the higher the likelihood of discoveries with low usefulness.
H₄: The more ordered/organized the body of materials being browsed, the higher the likelihood of discoveries with high usefulness.

I also think that organization trumps size, in that organization can mitigate the effect of size on degree of usefulness. Thus, browsing a very large, highly organized collection is likely to produce highly unexpected and highly useful serendipity.

Of course, another critical part of the relationship between browsing and serendipity is the browser herself. As Pasteur said “In the field of observation, chance favours only the prepared mind.”

Based on what I said above, I might expand/revise this to say: “Serendipity favors a highly prepared mind, browsing a large, well-ordered collection.”

Stay tuned … more later.