Bridging the gap: One view on what scientific methods have to offer to music research

Since the shift from largely structure-focused to more listener-focused music analysis, there has been increasing interest not only in questions of music perception and cognition, but also in methods borrowed from the sciences, in particular computational analysis and behavioral experimentation. In his third 1999 Ernest Bloch Lecture (, David Huron offers a survey of trends in research methodology that attempts to reconcile postmodernism and scientific empiricism. Focusing on the second part of this lecture, reflect on the following questions: According the Huron, what are the main (methodological) differences that arise in relation to these two perspectives? How do these differences relate to the two forms of skepticism he identifies? What is the “moral” dilemma that faces the researcher and how can this inform a researcher’s choice of methods? And finally, what are some of the advantages and disadvantages of applying empirical methods to music research questions?

In preparation for class discussion, please post a short preliminary comment, no later than Tuesday, September 3, 11:59 PM. You may respond to one of the above questions or raise other questions you find particularly interesting. Make sure to review your colleagues’ responses before class meets.



6 thoughts on “Bridging the gap: One view on what scientific methods have to offer to music research

  1. There is obviously a large amount to take in from this article, and so in the spirit of a brief initial response, I will focus on just a few points for now.

    From a purely practical standpoint, postmodernism (all interpretations are valid) and empiricism (which engages and makes newly available large sets of information) make the field more difficult to navigate, simply because there is so much more to take in. And in this light, it seems more important than ever to embrace a multitude of methodologies in order to contain this expansion.

    One way of dealing with such a burgeoning of possibilities is to trim away falsehood (where possible) and see what remains, but nothing in Huron’s discussion hints at how exactly we are to construct a tree of possibilities. This seems to be one of the greatest challenges. Without a comprehensive tree of possibilities, are we able to arrive at any conclusion of merit? Or will we always risk missing a potentially crucial factor?

    The rather daunting academic landscape created by such information influx points to one practical reason (in addition to those given by Huron) why empirical methods meet with such resistance in music scholarship. Empirical methods require scholars to learn an entirely new methodology—one that includes a significant expertise barrier; it is outside the comfort zone of many to learn an entirely new field mid-career. And in this sense, Kuhn’s notion of incommensurable paradigms seems relevant: perhaps we just have to wait for the new to replace the old.

    For any aspiring scholar, however, the re-education required to employ different methodologies depending on the situation at hand seems morally obligatory from the stance of good scholarship. Huron really convinced me of the value inherent in being able to tailor one’s methodology to the questions being asked. For instance, empirical study can verify the accuracy or inaccuracy of music theoretical claims (e.g., the melodic arch), but it cannot say anything about the human motivations that drive such trends. For that we need a different methodology.

    On a somewhat related note, I would also add that empirical studies can easily give the false impression of being objective when human agency is actually at play. In other words, we must really think about what empirical methods are actually capable of telling us. For instance, one might ask whether Mozart uses II6 or IV more often as a predominant chord in cadential progressions and have a computer run through a comprehensive data set to determine the answer. But that the act of harmonic analysis is itself an interpretive act adds a layer of complexity: the identity of a harmony is often ambiguous even when the function is clear.

    Well that was a rather disjointed response, but as I said, there is so much to talk about in this article. Hopefully these initial thoughts will spawn some interesting discussions.

  2. David Huron argues that, despite historical practices, the humanistic academy and the scientific academy should fundamentally differ by their subjects of study rather than their methodologies. He believes that empirical methods and postmodern discourse have a great deal of potential to interact symbiotically. Specifically, he advocates for the use of empirical and computational methods within musicology; this lecture can be seen as an attempt to validate future scholarship along Huron’s chosen career path.

    I found a lot to like in Huron’s discussion. New empiricism in music carves out a niche for voicing skepticism through inferential statistics. Though Huron is largely interested in vindicating theory-conserving skeptics and encouraging them to integrate statistics (in response to both denigration from scientists and the humanitarian impulse to avoid numbers), he allows for the converse, which I personally find the more appealing. Though “more likely” cannot equate “True,” significance can help justify an assertion that is seen as conventionally off the wall (as Huron puts it “Untested quackery has a [] chance of being helpful” {6}). OR, it can serve to discredit the dominant paradigm and question conventionality; I found his gap-fill analysis refreshing.

    However, there were quite a few points in the lecture with which I did not wholly agree or, at the very least, felt that his presentation proved somewhat troublesome. Here are five such items that troubled me (we will have much to discuss in class):

    (1)I do not understand his relationship with “deductive” methodologies. He references pure math as a traditional positivist pursuit but does not give us any sense of how this (a fundamentally important methodology in musicology historically) might interface with empirical methodologies. I found this an odd oversight because of his emphasis on what can and cannot be “proven.”

    (2)I feel that he misrepresents “retrospective data” in the context of historical musicology. As he posits it, Huron would be damning any theory based on an isolated data set as purely tautological.

    (3)His continual references to the “esthetic repercussions” are never developed. I have a sense that we should not sentence someone to life in jail on a preponderance of the evidence for the possible “moral” repercussions, but what might an esthetic consequence be?

    (4)I feel that his construction of the postmodernist is a bit of a straw man. By conflating the ideas of several generations of humanitarians, Huron’s caricature is concerned with morality beyond all reason that opposes all science and most forms of discourse. This postmodernist seemingly cannot believe anything — he must abandon all rules and distrust all scholarship (in Huron’s defense, he does warn his audience that he verges on over-simplification and is potentially breaking a “principal tenant of postmodernism” by attempting to represent their world-views [28]).

    (5)The y-axis (pitch height) in his melodic arch chart is unlabeled; 7.8 what I wonder? More seriously, he states that “averaging together thousands of melodic phrases constitutes the epitome of quantitative lunacy.” He is forthright; in a paper advocating inferential statistics, he should certainly have attempted a better interpretation of his data than simply averaging pitch heights. — We can easily create a situation that produces his precise chart without a single melodic arch!!! If we imagine a data set in which 3,182 melodies are of type A [10, 0, 0, 10, 0, 0, 10] and 3,182 melodies are of type B [5.6, 17.4, 17.8, 7.85, 17.6, 17.15, 4.8] (ordered sets of pitch heights), we would produce Huron’s result without the presence of a single arch. (Perhaps I am just a “theory-discarding skeptic” and need to abandon all faith in the humanities.)

    Despite the issues I had with Huron’s Lecture, I am largely on board with his project. I think that “empirical” research is only one possibility for quantitative methods, and I can imagine a future for the musicological disciplines that willingly embraces a variety of methodological practices from outside the humanities.

    • A thought on item (2)

      I had the same reaction at first based on this statement: “One of the most pernicious problems plaguing historical disciplines is the tendency to use a single data set both to generate and to support the theory.” And I also take issue with the implicit devaluation of
      much humanities scholarship.

      It then occurred to me, however, that retrospective data is a relative term (relative to the researcher), and so far as I can tell, very few humanities scholars are aware of *all* pertinent data to their studies. And even if they are aware, it is rare (in my experience) that theorists engage in a comprehensive sweep of available data before theorizing.

      It thus stands that there is usually at least some prospective data (the kind the researcher has not yet seen) that would allow for verification of a theory. And, of course, applications of theories to other unexpected (but related) domains also allows for verification.

  3. First, a caveat: my responses are synthetic. They are the product of my interpretation of only Huron’s text (1999) and my motivation to mold to the questions. In short, they are not the conclusions and convictions of my own study.

    What are the main (methodological) differences that arise in relation to these two perspectives?

    At bottom, the tacit methodology of postmodernism, if it may be said to possess one, is severe and single-minded reductionism. As Huron emphasizes: “everything reduces to politics”. Otherwise, via Huron, practitioners of postmodernism seem less attached to a particular methodology and more to a way of thinking: thinking reflexively about one’s subject position, about the symbolic and cultural milieus, and necessarily about power relationships. Huron’s postmodernism positions radically opposite from induction, the traditional bastion of scientific-empirical enquiry, and it appears to scotch one’s capacity to prefer one interpretation/theory to another. According to such thinking, calling another’s scholarship spew is a bald power grab, and perpetrator Taruskin, in terms of the postmodernism discussed here, could never be ideologically safe.

    Scientific empiricism, which arguably demonstrates a more subdued tendency toward incendiary of subversive undertones (or overtones?) and aspires to more dispassionate argumentation than the sometimes rigor-wanting waxing of the dispossessed, proceeds fundamentally through induction. ‘Induction’ as a concept, for long philosophically vexed, has sustained numerous patches, three of which Huron cites as influential to scientific thinking: falsification, conventionalism, and instrumentalism. The latter two constitute the dominant mode for pragmatic practitioners.

    How do these differences relate to the two forms of skepticism he identifies?

    Huron proposes a general equation, which intends to loosely characterize the two poles of institutionalized skepticism: the so-called ‘science/false-positive and humanities/false-negative association’. Huron suggests that the two realms of knowledge creation are not so opposed as is commonly conceived; both modes of inquiry are fundamentally motivated by skepsis, even if their respective versions stand in reciprocal relation. Yet neither skepticism is epistemological safer, and the researcher must always negotiate between the probabilities of false-positive and false-negative errors.

    Regarding the connection implied by the leading question: is not clear to me that Huron equates the humanities with postmodernism. Instead, Huron associates the two reciprocal forms of skepticism—theory-discarding and theory-conserving—with the sciences and the humanities, respectively. In other words, I am not confident enough to say that Huron argues that postmodernism exercises a theory-conserving skepticism.
    What is the “moral” dilemma that faces the researcher and how can this inform a researcher’s choice of methods?

    After the true sense of the word ‘dilemma’ must the research repeatedly negotiate between equally undesirable outcomes, in statistical terms, between the probability of false-positive error and false-negative error. According to Huron, the research is under a moral obligation to be in tune with the moral and esthetic consequences of committing one error or the other. The situation will suggest the balance, but decisions can ultimately be trying. High-risk fields, such as medicine, will call for a different calibration of the balance than low-risk, abstract, and impersonal ones.

    And finally, what are some of the advantages and disadvantages of applying empirical methods to music research questions?

    Huron is clear in his support for a more catholic appreciation and application of quantitative methods in music scholarship. In his advocacy he apes some of the common gripes while proposing several compelling advantages. Huron argues that quantitative approaches allow the researcher to observe patterns of organization that might otherwise be difficult to discern. In other words, a state-of-the-art computer and software could turn a slogging theory adjunct through hi-tech prosthesis into a sort of Allen Forte Robocop. That one could possess such super-human faculties raises a whole host of questions—ideological to formal to aesthetic—which are surely beyond the scope of this response, but which I also note seem irrelevant to Huron. For the latter says only that the prosthesis allows one “to become a more observant music scholar”. If quantitative methods could facilitate one’s musical bildung, then surely there is some merit… Another virtue of quantitative methods is that they may confirm, refute, or even dissolve some or many of the most long-standing musical intuitions handed down. Huron zeros in on the ‘melodic arch’ and the ‘gap-fill’ technique—incidentally (or not), both were given due attention by Meyer’s Penn School—, demonstrating how received wisdom may be validated in the former case, or rewritten in the latter. (The ‘gap-fill’ as a musical concept has no quantitative reality.)

    One of the more obvious disadvantages of the quantitative method is its dehumanizing reputation, namely, forcing phenomena into numerical categories that may not be appropriate. Here, Huron spends some time in subtle argument, calling to aid the inviolate Socrates. In closing, I admit to being impressed in part by Huron’s consideration of the disadvantages of a quantitative approach. Specifically, he can imagine the moral hazard of quant practitioners toward reification and fetishism: numbers are king, musical intuitions are folk tales, etc.

    -S P G

  4. Reposting this from my original response…

    David Huron (1999) raises an interesting point regarding differences in training between fields. Scientists often are not exposed to theory of knowledge, and humanities folk are often blind to research methodologies (at least with regards to more empirical aims). Having been in a graduate program for both fields (as a masters student in music theory and later in psychology), I find that this statement could not be more true. My friend (who also happens to be a PhD student in musicology, with a masters in philosophy) always argued that my scientific (and possibly, music theoretical, as well) background blinded me with regards to what knowledge is. He mocked the fact that many scientists receive a PhD but know nothing about the philosophical underpinnings of their research, and argued that we should all be required to take seminars, at least, on Hume. All of our arguments would always end up in the same place: Him asking me “But, what is knowledge?” and my frustrated eye roll. Later, I would laugh about it with my psychology lambastes. But never did we (my friend, myself, or my labmates) actually sit back to evaluate each other’s points of view, each other’s research goals and questions, or assess the validity of the other field’s truth-seeking. I must admit that as masters students in music theory, we also were never exposed to these larger questions mainly due to the segregation of the musicologists and theorists, especially with respect to seminar choices. Not until I arrived at Yale, have I seen such an integration of scholars, whether it be amongst faculty or seminar-student distribution. Perhaps this is due to the type of scholarship that arises from state schools versus more liberal arts-minded humanities programs. But I digress…

    While working to bridge the two fields in my own research, I find the question of knowledge and attainment of truth is increasingly becoming a huge hurdle. I believe, like Huron mentions, in the biological aspects of humanity: without our highly developed cortex, we are just animals largely directed by instinct and disposition toward evolutionary survival. We would never have come up with philosophical thought in the first place (I don’t know, maybe my dog thinks about truth all the time but just has no way to communicate that with me?). In either case, somehow our physiological and biological activities have led to the development of culture, language, and art, all tied into abstract concepts of emotion, thought, social context, and so forth. Music is therefore not just a complex set of perceptual inputs, but also a process that involves agency, emotion, and social context. As a theorist, I could default to our antiquated method of scholarship, analyzing music without regard to social context; there’s certainly still a community that would accept that research with open arms. As someone working with perception, I could default to low-level research: this is how my brain works in perceiving sound waves, and that’s all music is. But what about the evidence and knowledge I have on what music is? What is the point of undertaking perceptual, analytical, and historical research without wanting to find a real truth, something that fits the evidence from all points of view? As Huron says, postmodernism usually leads us to the same conclusion, if there is not true knowledge, then we are at standstill since we can never attain it. But I am not one to give up.

    Solving this problem, the problem of music, from either a cognitive, historical, theoretical, or any other point of view then should include multi-faceted methodological techniques (I look forward to reading Huron’s proposal as I near the end of the lecture). However, a problem then arises, that of educating ourselves to understand as many forms of attaining truth (or attempts at attaining truth) as we can. We can only take so many seminars in two years of course work; perhaps I should take that Hume seminar after all…

Leave a Reply

Your email address will not be published. Required fields are marked *