I was genuinely surprised recently when someone told me they didn’t believe there was a real distinction between fundamental and applied work. They asked, “What do you mean by fundamental? You keep implying that some of us don’t do it.” I tried to explain that “fundamental” does not mean “superior,” and that my point wasn’t a criticism but simply a statement about the type of science we were discussing. I wasn’t successful in convincing them, so I’m writing this post in the hope that it will clarify the distinction for students and for anyone else who may share that misconception.
The distinction between fundamental/basic and applied science has a long history. While I'm sure we can find some references to such distinctions in ancient philosophy, most people trace the distinction back to Vannevar Bush, who made the distinction when coming up with policy advice for the US president. This article, for example, describes that distinction and identifies an issue with how "basic/fundamental" and "applied" are defined. They argue that the definitions are rooted in the researcher's intentions, and because such intentions do not matter in the end, the distinction is useless. They also construct what I think is a strawman and argue against this sort of a linear approach to doing science:
Basic Research–>Applied Research–>Product Development–>Market Diffusion and Deployment.
Of course, all of us understand that discovery and innovation go hand-in-hand. Does that mean there is no distinction between fundamental and applied work? Absolutely not.
Fundamental vs. Applied Science
Fundamental Science: Definitions
Let us first examine an example from physics. Newton came up with the laws of motion, not just based on observations but also based on a lot of theoretical reasoning for which he had to invent an entire new branch of mathematics (calculus). What this allowed him to do was not just provide a coherent framework that could explain all existing observations but also one that could also make predictions about unseen scenarios. Contrast that with the several wonderful observations that Galileo made about the moon, planets, objects that stayed in motion on an inclined plane, etc. These observations played an important role in Newton's theory, but would we say that Galileo did fundamental/basic science? Perhaps Vannevar Bush might, because Galileo's intentions in making these observations were driven by curiosity and not any particular use case. However, it is clear that these were just observations that did not immediately evoke any explanations for the fundamental mechanisms by which they came about. That work was done by others. Observations that come without precise, testable explanations do not constitute fundamental science.
References to such a notion of fundamental science can be found all the way back in
Francis Bacon's 17th century work. The following quote from the link in the previous sentence is worth reading in its entirety:
“Bacon's antipathy to simple enumeration as the universal method of science derived, first of all, from his preference for theories that deal with interior physical causes, which are not immediately observable” (Urbach 1987, 30; see: sect. 2). The last type can be supplemented by tables of counter-instances, which may suggest experiments:
To move from the sensible to the real requires the correction of the senses, the tables of natural history, the abstraction of propositions and the induction of notions. In other words, the full carrying out of the inductive method is needed. (Malherbe 1996, 85)
The sequence of methodical steps does not, however, end here, because Bacon assumes that from lower axioms more general ones can be derived (by induction). The complete process must be understood as the joining of the parts into a systematic chain. From the more general axioms Bacon strives to reach more fundamental laws of nature (knowledge of forms), which lead to practical deductions as new experiments or works (IV, 24–5). The decisive instruments in this process are the middle or ‘living axioms,’ which mediate between particulars and general axioms. For Bacon, induction can only be efficient if it is eliminative by exclusion, which goes beyond the remit of induction by simple enumeration. The inductive method helps the human mind to find a way to ascertain truthful knowledge.
Therefore, fundamental science is not just the act of gathering observations and is not even about deriving laws if the laws are just mere summaries of observations (more about this in the next section). The act of doing fundamental science involves coming up with "theories that deal with interior physical causes" of the observations. Is the act of gathering information important? Absolutely. Does it constitute fundamental science? Absolutely not.
Examples from cognitive science
In the late 19th century, experimental psychologists such as Hermann Ebbinghaus and Edward Thorndike began collecting systematic, quantitative data on learning and memory. This marked a significant move away from purely introspective methods toward making psychology an empirical, laboratory-based science, a transition later accelerated by early 20th-century behaviorism. Let me use the specific example of Ebbinghaus' work on forgetting curves to illustrate why gathering observations and even deriving a law out of them does not constitute "fundamental" science.
In 1885, Hermann Ebbinghaus conducted months of self-experiments to study memory using lists of meaningless “nonsense syllables” (e.g., BOK, ZAT). He memorized lists until he could recite them perfectly twice in a row, then tested himself after varying time intervals. By comparing the repetitions needed to relearn the list with the original effort, he calculated “savings” in relearning. His results revealed the famous forgetting curve (Fig 1), which captured a rapid loss of memory within the first hours or days, followed by a slower decline.
Fig 1. Ebbinghaus' forgetting curve.
It wasn't until
much later than people fit mathematical functions to such data and tried to offer theoretical explanations of how such mathematical functions of forgetting could arise. However, fitting mathematical functions to a century of observational data still does not count as fundamental science in my view. They are precursors to fundamental theoretical work that proposes the
internal physical causes of forgetting.
Therefore, there is a clear conceptual difference between descriptive laws and causal/mechanistic theories. The sort of work that was the topic of the discussion that I mentioned at the beginning of this post did not even attempt to offer systematization of observations through even descriptive laws, let alone mechanistic theories. The sort of work that I was referring to involved mere data collection and exploratory reports of whatever patterns and correlations came up in the data. This is what they argued was also "fundamental." Sorry, but I strongly disagree for all the above reasons.
In contrast to efforts to fit descriptive mathematical functions, others have tried to offer mechanistic explanations of forgetting. For example, a
theory that guides a lot of thinking in our own lab proposes that forgetting arises because the mental/contextual state associated with encoding drifts over time, making retrieval cues less effective. The theory is stated in precise mathematical terms. We can simulate memory behavior using the theory. We can infer model parameters from behavioral data and try to offer mechanistic explanations of the behavioral data. Therefore, Howard and Kahana (2002) did the hard fundamental theoretical work.
However, all we do really in the lab is test the theory through designing experiments or doing analyses of secondary datasets. So I would argue that even what we do ourselves doesn't constitute true fundamental work in cognitive science, though with an intentional effort to test specific predictions of the theory, that sort of work falls into the middle of the spectrum between fundamental and applied work. So, I actually agree with Vannevar Bush that the researcher's intentions also contribute to whether the work is fundamental or applied, even if there is certainly a spectrum between the two. Just because there exists a spectrum does not mean there is no meaningful distinction between the ends of the spectrum! The existence of a spectrum also certainly does not argue against the utility of the fundamental and applied labels. These labels help guide policy, just as Vannevar Bush suggested. The same framework can and should guide not just policy at the national level but even at the level of departments. For example, it really depends on a department's vision for its programs and research portfolio whether they should hire folks who do work that is closer to the fundamental or applied side of the aforementioned spectrum.
Distinctions between fundamental science and other kinds of observational or applied work have existed for centuries, and for good reason. If you want to argue against these distinctions, the burden is on you to clearly show why removing them is justified and how doing so would lead to meaningful practical benefits.
Comments
Post a Comment