Between Natural Sciences and Social Sciences/Humanities---Methodologies, Goals, Gaps, Divides and Complements*
by Shuo Zhang
APR.25,2011
Max Weber thought that "statements of fact are one thing, statements of value another, and any confusing of the two is impermissible." Ralf Dahrendorf writes in his essay "Max Weber and Modern Social Science", acknowledging that Weber clarified the difference between pronouncements of fact and of value (Max Weber's View of Objectivity in Social Science by Steve Hoenisch, criticism.com). I remember one computational logician at a SELLC 2010 banquet remarked that social scientists, such as linguists and philosophers, rely heavily on subjectivity, hunch, reasoning, subjective responses (survey), even guesses, values, opinions, etc., while natural or computer scientists rely on facts, experiments, observations, quantitative results, reproducibility, etc. While it is debatable today how linguists have adopted methods more like a hard science (like experimental phonetics), it is generally accepted that there are huge differences in the ways these disciplines work to develop their theories and research, like the lines quoted about Weber and his fellow thinkers at the beginning of this paragraph. In this essay, however, I am going to approach this issue from a different angle, a rather practical one. The arguments I'm going to make lie deep in my experiences in dealing with subject matters of human behavior (music and language) from various perspectives, and a big part of it stems from a particular discussion, when I talked at length about my quasi-scientific comprehensive review papers of speech and singing with my committee of music professors at Pitt music dept(which are of great help to me in thinking about issues in this article). In this article,I will show how a deeper appreciation of these different approaches and a more encompassing perspective will benefit our research. Meanwhile, the profound implications of their differences are also discussed, in the light of the most recent research literature in linguistics.
It is generally acknowledged that the "hard" sciences and social sciences/humanities (since humanities is more at the opposite end of the spectrum against hard sciences, to the purpose of this essay, I will use humanities for discussions-which also apply to those social sciences such as literary studies, political science, history and cultural anthropology- and leave out the term social sciences from now on, although a lot of times you will see I have particularly musicology/ethnomusicology in mind) have different methodologies, one through strictly controlled experiment and calculation, the other through more subjective analysis, investigation, fieldwork, interviews, etc. But not everyone is aware of the usual misunderstandings of how scholars in the two camps view each other. Although I have sit through numerous seminars in musicology, I used to assume that humanities scholars accept as a fact that their work is more subjective and is less precise and accurate thus credible than hard sciences (considering a lot of what they deal with involve more or less opnions and values, no matter their own, or others). But I realized that this impression is false only very recently (despite my Music Analysis Seminar professor used to tell us "music is not a science"). I will show this conception in the next few paragraphs.
It is true that humanities research appear to be more subjective on many levels. To put it in plain words, their studies usually involve one telling a story (such as in ethno/musicology) and analyzing the various phenomena/processes/mechanisms/forces behind the story. This constitutes an essential part of the study, while massive attempts to build "theories" are generally less credible/favored due to the fact that theories are more case-specific, and the application to other cases usually involve subjective processes of reasoning (consider, for instance, the research question "What do the unprecedented changes in erhu music landscape indicate in this period? How do we view this phenomenon in the greater context of China in the 1990s and in the new millennium?"(Zhang 2009)) Debates in seminars usually lead to no one correct conclusion, and the point of any scholarly discussion is certainly not to reach that conclusion either---but merely to interact, interpret, and discuss the different perspectives. That is the beauty of the process. The goal of the debate is not to reach a universally correct "truth".
We may think from this that hard sciences reveal more "truth" than humanities, truths that can apply more universally and be put into practice in a clear-cut fashion(consider when a computer program runs correctly, you get things done, as opposed to when it runs incorrectly, then you don't). But that is not what it looks like in the minds of the humanities scholars, especially when dealing with subject matters like music or language, the behavioral or cognitive studies of the mind. In fact, in the course of the history of ethnomusicology, there was a period of time where machine transcriptions(melograph) and analysis of musical sound were available and were utilized. Linguistic models of analysis also were borrowed and widely applied. But people soon came to realize that these do not fit the complexity of the subject matter--they reveal very little about the numerous different ways people listen to, enjoy, respond to, and experience music. We human beings have rich and subjective minds, and even in a seminar of world music transcription and analysis (in a Ph.D program in ethnomusicology), the goal, according to the instructor, is clearly not to sit through the seminar so that you can be better at transcribing music--but just to put forth the various ways people experience music.
Today, the advancement of music cognition is flourishing more than ever. As far as the bigger picture of the discipline goes, while we sometimes criticize that the research rely on Western art music too heavily, we should also speculate the implications of history: Why did machines and linguistic models of analysis fade out from the study of world music?How are the approaches of music cognition today different from then? And more importantly, what do we have to do in order to bring ethnomusicology on board into the game of music cognition again?(as many scholars, such as Ian Cross and David Huron have called for)
Let's revisit the idea from above that humanities scholars do not think that sciences necessarily (let's focus on the case of music and language cognition) reveal more "truths" than their research in the humanities. How so? Fundamentally speaking, and in a nutshell, they do not believe that a series of numbers and charts can tell the whole truth of what is behind the vastly complicated phenomena of human mental activities, in this case, music, or language. (Similar debates exist between linguistic anthropology and linguistics, where the former criticizes the latter being too simple-minded and essentialist to regard language in a purely logical and formal way-formal as in formalism, which essentially deprives context from the subject matter). In their minds, the results from a few experiments are too oversimplified and reduced (in a sense of reductionism) from the densely layered and diversely rich truths. They hold this principle dear to their heart, even if we would argue that sciences use "empirical experiments" to prove things with evidences while they do not.
Is that so? If you think in a take-for-granted fashion that science always reveals more "truths", think again. Consider this example. Patel and Xu (2010) revealed in an experiment that Mandarin Chinese preserves over 90% of intelligibility in a non-noisy background. If we were to conclude from this just how tone languages work in terms of the robustness of our perception, we would have probably gone too far. Would we? Well, Bell Yung of Pitt Music Department suggests that MC is a language probably with more robustness than other tone languages. Why? MC is the administrative and official language of China, where hundreds of millions of people would speak it since childhood but with a considerable amount of dialects, often tone variations. Therefore an average adult MC speaker would have heard a large varieties of phonetic, especially tonal performances in their life. If we do this experiment again with a less widely spoken language, then it probably would be different. Or not. The results are not important as the idea--where cultural factors come into play. And sometimes science has to take into account those various cultural factors, which they in general do not like--cultural factors that are irregular, unsystematic,harder to observe to derive a pattern or rule, and harder to represent using quantitative models.(Although in this case, according to the recent research, including my own, I have to say that language communication is much more robust than we usually assume. So Dr.Yung has a valid point but the result may not be that different--we'll see).
On the other hand, in terms of the complexity of the story, science has very different goals than humanities. It is well known, for example, that physicists are trying to find a more unifying and simple explanation for all the forces in the universe. It is a goal set up since before the time of Albert Einstein and other great minds in their field, that the ultimate truth must be expressed in an incredibly simple, (maybe symmetric), elegant, and clear equation, just like E=mc^2. If a equation is too complicated, it is ugly. They believe that is how god has intended it--the truth is simple and clear, and therefore elegant.
This is certainly not the case in humanities. In the case of ethnomusicology, for instance, people constantly criticize that concise theoretical models of viewing things are too simplistic, not sufficient to capture the depth and the complexity of the phenomena in question. The use of sets of dichotomies, for example, has been criticized over and over again and replaced by detailed ethnographies--scholars constantly argue that the truth is not cut-and-dry, and cannot be summarized with pairs of dichotomies, such as tradition/modernity, globalization/localization, hybridity/indigenization, to name a few. Ethnomusicologists would argue, for instance, that viewing musical cultures by the tradition/modernity dichotomy is problematic, because tradition is not static, it keeps changing, keeps evolving, vary by individual's perception, and what is "traditional" and "modern" are constantly mixing to create new bodies of hybridity. It is by no means as clear as 0s and 1s or positive and negative integers, and it is therefore problematic to view this in a simple dichotomy. By the same token, "modernization" is theoretically supposed to be "replacing" the tradition while in reality they intermingle with them, and there are constant counter-hegemony forces and voices in society-it is never simple. They would even go as far to deny the usage of dichotomy between music and language--by arguing that what we call "music" or "language" isn't necessarily as clearly defined in Amazonian forest tribal cultures (some vocal performing genres are considered neither music nor language by the locals, but something else)(Seeger 2004). Therefore you're never encouraged to talk about things in a simple model, for that deprives the rich texture of what really happens--an essential feature of story telling. If you strip down the stories into a few points, there is no point in telling it.
How then, will the mix of these two perspectives benefit our research and improves our understanding of human behavior? Or will it? The implications of the existence of these two fundamentally different approaches and perspectives are profound. (1)On the one hand, it seems to me that science can definitely benefit from taking into account the cultural factors (which is exactly why there have always been scholars calling for a bridge between sciences and humanities), such as in the case suggested above (the examples of the robustness of tone language perception). Only that way can we get closer to a wholesome understanding of the phenomena in human behavior. (2)On the other hand, the two divide on important matters such as the role of culture and brain/biology in the development of human music and language (and other behavioral patterns or products). Humanitistic approaches tend to criticize science's tendency to "over generalize"(example: Patel and colleagues's work on how speech rhythm is reflected in music rhythm of that particular culture (2006), or Deutsch's work on how one's musical absolute pitch is related to their native language-tone or non-tone language(2009)). While some of the controversies in (2) can be resolved to the scenario in (1) in an attempt to reconsider more data and derive more credible conclusions, others might go on for a long time as a topic of debate. Recent developments in the field, however, also revealed new ways of resolving the two, as we see in the media reports of a study (using mathematical models) on the evolution of language (originally published in Nature):
"A long-standing idea that human languages share universal features that are dictated by human brain structure has been cast into doubt. A study reported in Nature has borrowed methods from evolutionary biology to trace the development of grammar in several language families. The results suggest that features shared across language families evolved independently in each lineage. The authors say cultural evolution, not the brain, drives language development. At the heart of both studies is a method based on what are known as phylogenetic studies....That is inconsistent with the dominant 'universality theories' of grammar; it suggests rather that language is part of not a specialised module distinct from the rest of cognition, but more part of broad human cognitive skills....The paper asserts instead that "cultural evolution is the primary factor that determines linguistic structure, with the current state of a linguistic system shaping and constraining future states...."We're not saying that biology is irrelevant - of course it's not," Professor Gray told BBC News. "But the clumsy argument about an innate structure of the human mind imposing these kind of 'universals' that we've seen in cognitive science for such a long time just isn't tenable.""(Media reports from BBC, original article published in Nature/full article here)
REFERENCE
Deutsch, D., Dooley, K., Henthorn, T. and Head, B. Absolute pitch among students in an American music conservatory: Association with tone language fluency. Journal of the Acoustical Society of America, 2009, 125, 2398-2403.
Dunn, M., Greenhill, S. J., Levinson, S. C., & Gray, R. D. (2011). Evolved structure of language shows lineage-specific trends in word-order universals. Nature. Advance online publication. doi:10.1038/nature09923.
Patel, A.D., Xu, Y., & Wang, B. (2010). The role of F0 variation in the intelligibility of Mandarin sentences. Proceedings of Speech Prosody 2010, May 11-14, 2010, Chicago, IL, USA. (paper-pdf)
Patel, A.D., Iversen, J.R., & Rosenberg, J.C. (2006). Comparing the rhythm and melody of speech and music: The case of British English and French. Journal of the Acoustical Society of America, 119:3034-3047. (paper-pdf) (popular version of paper-html) (Nature news story-html) (online nPVI calculator)
Seeger, A. Why Suya Sing? Champaign, IL:University of Illinois Press, 2004.
Zhang,S. Erhu as Violin: The Development of the Chinese Two-Stringed Bowed Lute, c.1990-2008. MA THESIS:Department of Music, University of Pittsburgh, 2009.
*Thank you for reviewing this paper/past discussions to the following research scientists/professors: Ani Patel, Ian Cross, Bell Yung, David Huron, Andrew Weintraub, Adriana Helbig,David Lightfoot,Donna Lardiere, Elizabeth Zsiga.
It is generally acknowledged that the "hard" sciences and social sciences/humanities (since humanities is more at the opposite end of the spectrum against hard sciences, to the purpose of this essay, I will use humanities for discussions-which also apply to those social sciences such as literary studies, political science, history and cultural anthropology- and leave out the term social sciences from now on, although a lot of times you will see I have particularly musicology/ethnomusicology in mind) have different methodologies, one through strictly controlled experiment and calculation, the other through more subjective analysis, investigation, fieldwork, interviews, etc. But not everyone is aware of the usual misunderstandings of how scholars in the two camps view each other. Although I have sit through numerous seminars in musicology, I used to assume that humanities scholars accept as a fact that their work is more subjective and is less precise and accurate thus credible than hard sciences (considering a lot of what they deal with involve more or less opnions and values, no matter their own, or others). But I realized that this impression is false only very recently (despite my Music Analysis Seminar professor used to tell us "music is not a science"). I will show this conception in the next few paragraphs.
It is true that humanities research appear to be more subjective on many levels. To put it in plain words, their studies usually involve one telling a story (such as in ethno/musicology) and analyzing the various phenomena/processes/mechanisms/forces behind the story. This constitutes an essential part of the study, while massive attempts to build "theories" are generally less credible/favored due to the fact that theories are more case-specific, and the application to other cases usually involve subjective processes of reasoning (consider, for instance, the research question "What do the unprecedented changes in erhu music landscape indicate in this period? How do we view this phenomenon in the greater context of China in the 1990s and in the new millennium?"(Zhang 2009)) Debates in seminars usually lead to no one correct conclusion, and the point of any scholarly discussion is certainly not to reach that conclusion either---but merely to interact, interpret, and discuss the different perspectives. That is the beauty of the process. The goal of the debate is not to reach a universally correct "truth".
We may think from this that hard sciences reveal more "truth" than humanities, truths that can apply more universally and be put into practice in a clear-cut fashion(consider when a computer program runs correctly, you get things done, as opposed to when it runs incorrectly, then you don't). But that is not what it looks like in the minds of the humanities scholars, especially when dealing with subject matters like music or language, the behavioral or cognitive studies of the mind. In fact, in the course of the history of ethnomusicology, there was a period of time where machine transcriptions(melograph) and analysis of musical sound were available and were utilized. Linguistic models of analysis also were borrowed and widely applied. But people soon came to realize that these do not fit the complexity of the subject matter--they reveal very little about the numerous different ways people listen to, enjoy, respond to, and experience music. We human beings have rich and subjective minds, and even in a seminar of world music transcription and analysis (in a Ph.D program in ethnomusicology), the goal, according to the instructor, is clearly not to sit through the seminar so that you can be better at transcribing music--but just to put forth the various ways people experience music.
Today, the advancement of music cognition is flourishing more than ever. As far as the bigger picture of the discipline goes, while we sometimes criticize that the research rely on Western art music too heavily, we should also speculate the implications of history: Why did machines and linguistic models of analysis fade out from the study of world music?How are the approaches of music cognition today different from then? And more importantly, what do we have to do in order to bring ethnomusicology on board into the game of music cognition again?(as many scholars, such as Ian Cross and David Huron have called for)
Let's revisit the idea from above that humanities scholars do not think that sciences necessarily (let's focus on the case of music and language cognition) reveal more "truths" than their research in the humanities. How so? Fundamentally speaking, and in a nutshell, they do not believe that a series of numbers and charts can tell the whole truth of what is behind the vastly complicated phenomena of human mental activities, in this case, music, or language. (Similar debates exist between linguistic anthropology and linguistics, where the former criticizes the latter being too simple-minded and essentialist to regard language in a purely logical and formal way-formal as in formalism, which essentially deprives context from the subject matter). In their minds, the results from a few experiments are too oversimplified and reduced (in a sense of reductionism) from the densely layered and diversely rich truths. They hold this principle dear to their heart, even if we would argue that sciences use "empirical experiments" to prove things with evidences while they do not.
Is that so? If you think in a take-for-granted fashion that science always reveals more "truths", think again. Consider this example. Patel and Xu (2010) revealed in an experiment that Mandarin Chinese preserves over 90% of intelligibility in a non-noisy background. If we were to conclude from this just how tone languages work in terms of the robustness of our perception, we would have probably gone too far. Would we? Well, Bell Yung of Pitt Music Department suggests that MC is a language probably with more robustness than other tone languages. Why? MC is the administrative and official language of China, where hundreds of millions of people would speak it since childhood but with a considerable amount of dialects, often tone variations. Therefore an average adult MC speaker would have heard a large varieties of phonetic, especially tonal performances in their life. If we do this experiment again with a less widely spoken language, then it probably would be different. Or not. The results are not important as the idea--where cultural factors come into play. And sometimes science has to take into account those various cultural factors, which they in general do not like--cultural factors that are irregular, unsystematic,harder to observe to derive a pattern or rule, and harder to represent using quantitative models.(Although in this case, according to the recent research, including my own, I have to say that language communication is much more robust than we usually assume. So Dr.Yung has a valid point but the result may not be that different--we'll see).
On the other hand, in terms of the complexity of the story, science has very different goals than humanities. It is well known, for example, that physicists are trying to find a more unifying and simple explanation for all the forces in the universe. It is a goal set up since before the time of Albert Einstein and other great minds in their field, that the ultimate truth must be expressed in an incredibly simple, (maybe symmetric), elegant, and clear equation, just like E=mc^2. If a equation is too complicated, it is ugly. They believe that is how god has intended it--the truth is simple and clear, and therefore elegant.
This is certainly not the case in humanities. In the case of ethnomusicology, for instance, people constantly criticize that concise theoretical models of viewing things are too simplistic, not sufficient to capture the depth and the complexity of the phenomena in question. The use of sets of dichotomies, for example, has been criticized over and over again and replaced by detailed ethnographies--scholars constantly argue that the truth is not cut-and-dry, and cannot be summarized with pairs of dichotomies, such as tradition/modernity, globalization/localization, hybridity/indigenization, to name a few. Ethnomusicologists would argue, for instance, that viewing musical cultures by the tradition/modernity dichotomy is problematic, because tradition is not static, it keeps changing, keeps evolving, vary by individual's perception, and what is "traditional" and "modern" are constantly mixing to create new bodies of hybridity. It is by no means as clear as 0s and 1s or positive and negative integers, and it is therefore problematic to view this in a simple dichotomy. By the same token, "modernization" is theoretically supposed to be "replacing" the tradition while in reality they intermingle with them, and there are constant counter-hegemony forces and voices in society-it is never simple. They would even go as far to deny the usage of dichotomy between music and language--by arguing that what we call "music" or "language" isn't necessarily as clearly defined in Amazonian forest tribal cultures (some vocal performing genres are considered neither music nor language by the locals, but something else)(Seeger 2004). Therefore you're never encouraged to talk about things in a simple model, for that deprives the rich texture of what really happens--an essential feature of story telling. If you strip down the stories into a few points, there is no point in telling it.
How then, will the mix of these two perspectives benefit our research and improves our understanding of human behavior? Or will it? The implications of the existence of these two fundamentally different approaches and perspectives are profound. (1)On the one hand, it seems to me that science can definitely benefit from taking into account the cultural factors (which is exactly why there have always been scholars calling for a bridge between sciences and humanities), such as in the case suggested above (the examples of the robustness of tone language perception). Only that way can we get closer to a wholesome understanding of the phenomena in human behavior. (2)On the other hand, the two divide on important matters such as the role of culture and brain/biology in the development of human music and language (and other behavioral patterns or products). Humanitistic approaches tend to criticize science's tendency to "over generalize"(example: Patel and colleagues's work on how speech rhythm is reflected in music rhythm of that particular culture (2006), or Deutsch's work on how one's musical absolute pitch is related to their native language-tone or non-tone language(2009)). While some of the controversies in (2) can be resolved to the scenario in (1) in an attempt to reconsider more data and derive more credible conclusions, others might go on for a long time as a topic of debate. Recent developments in the field, however, also revealed new ways of resolving the two, as we see in the media reports of a study (using mathematical models) on the evolution of language (originally published in Nature):
"A long-standing idea that human languages share universal features that are dictated by human brain structure has been cast into doubt. A study reported in Nature has borrowed methods from evolutionary biology to trace the development of grammar in several language families. The results suggest that features shared across language families evolved independently in each lineage. The authors say cultural evolution, not the brain, drives language development. At the heart of both studies is a method based on what are known as phylogenetic studies....That is inconsistent with the dominant 'universality theories' of grammar; it suggests rather that language is part of not a specialised module distinct from the rest of cognition, but more part of broad human cognitive skills....The paper asserts instead that "cultural evolution is the primary factor that determines linguistic structure, with the current state of a linguistic system shaping and constraining future states...."We're not saying that biology is irrelevant - of course it's not," Professor Gray told BBC News. "But the clumsy argument about an innate structure of the human mind imposing these kind of 'universals' that we've seen in cognitive science for such a long time just isn't tenable.""(Media reports from BBC, original article published in Nature/full article here)
REFERENCE
Deutsch, D., Dooley, K., Henthorn, T. and Head, B. Absolute pitch among students in an American music conservatory: Association with tone language fluency. Journal of the Acoustical Society of America, 2009, 125, 2398-2403.
Dunn, M., Greenhill, S. J., Levinson, S. C., & Gray, R. D. (2011). Evolved structure of language shows lineage-specific trends in word-order universals. Nature. Advance online publication. doi:10.1038/nature09923.
Patel, A.D., Xu, Y., & Wang, B. (2010). The role of F0 variation in the intelligibility of Mandarin sentences. Proceedings of Speech Prosody 2010, May 11-14, 2010, Chicago, IL, USA. (paper-pdf)
Patel, A.D., Iversen, J.R., & Rosenberg, J.C. (2006). Comparing the rhythm and melody of speech and music: The case of British English and French. Journal of the Acoustical Society of America, 119:3034-3047. (paper-pdf) (popular version of paper-html) (Nature news story-html) (online nPVI calculator)
Seeger, A. Why Suya Sing? Champaign, IL:University of Illinois Press, 2004.
Zhang,S. Erhu as Violin: The Development of the Chinese Two-Stringed Bowed Lute, c.1990-2008. MA THESIS:Department of Music, University of Pittsburgh, 2009.
*Thank you for reviewing this paper/past discussions to the following research scientists/professors: Ani Patel, Ian Cross, Bell Yung, David Huron, Andrew Weintraub, Adriana Helbig,David Lightfoot,Donna Lardiere, Elizabeth Zsiga.