Mentors, paper referees and grant reviewers have warned me on occasion about scientific "fishing expeditions," the conduct of empirical research that does not test a specific hypothesis or is not guided by theory. Such "blind empiricism" was said to be unscientific, to waste time and produce useless data. Although I have never been completely convinced of the hazards of fishing, I now reject them outright, with a few reservations.
I'm not advocating the collection of random facts, but the use of broad-based descriptive studies to learn what to study and how to study it. Those who fish learn where the fish are, their species, number and habits. Without the guidance of preliminary descriptive studies, hypothesis testing can be inefficient and misguided. Hypothesis testing is a powerful means of rejecting error — of trimming the dead limbs from the scientific tree — but it does not generate hypotheses or signify which are worthy of test. I'll provide two examples from my experience.
In graduate school, I became intrigued with neuroembryology and wanted to introduce it to developmental psychology, a discipline that essentially starts at birth. My dissertation was a fishing expedition that described embryonic behavior and its neurophysiological mechanism. I was exploring uncharted waters and sought advice by observing the ultimate expert, the embryo. In this and related work, I discovered that prenatal movement is the product of seizure-like discharges in the spinal cord (not the brain), that the spinal discharges occurred spontaneously (not a response to sensory stimuli), that the function of movement was to sculpt joints (not to shape postnatal behavior such walking), and to regulate the number of motorneurons. Remarkable!
But decades later, this and similar work is largely unknown to developmental psychologists who have no category for it. The traditional psychological specialties of perception, learning, memory, motivation and the like, are not relevant during most of the prenatal period. The finding that embryos are profoundly unpsychological beings guided by unique developmental priorities and processes is not appreciated by theory-driven developmental psychologists. When the fishing expedition indicates that there is no appropriate spot in the scientific filing cabinet, it may be time to add another drawer.
Years later and unrepentant, I embarked on a new fishing expedition, this time in pursuit of the human universal of laughter — what it is, when we do it, and what it means. In the spirit of my embryonic research, I wanted the expert to define my agenda—a laughing person. Explorations about research funding with administrators at a federal agency were unpromising. One linguist patiently explained that my project "had no obvious implications for any of the major theoretical issues in linguistics." Another, a speech scientist, noted that "laughter isn't speech, and therefore had no relevance to my agency's mission."
Ultimately, this atheoretical and largely descriptive work provided many surprises and counterintuitive findings. For example, laughter, like crying, is not consciously controlled, contrary to literature suggesting that we speak ha-ha as we would choose a word in speech. Most laughter is not a response to humor. Laughter and speech are controlled by different brain mechanisms, with speech dominating laughter. Contagious laughter is the product of neurologically programmed social behavior. Contrasts between chimpanzee and human laughter reveal why chimpanzees can't talk (inadequate breath control), and the evolutionary event necessary for the selection for human speech (bipedality).
Whether embryonic behavior or laughter, fishing expeditions guided me down the appropriate empirical path, provided unanticipated insights, and prevented flights of theoretical fancy. Contrary to lifelong advice, when planning a new research project, I always start by going fishing.