In the context of the Fischer reading for this past week, I would like to take another look at Ben Singer's article "Manhattan Nickelodeons: New Data on Audiences and Exhibitors."
In the article, Singer presents new data on the class compositions of four Manhattan neighborhoods – Little Italy, Jewish Harlem, Yorkville, and Harlem proper – designed to challenge Robert Allen's argument that many nickelodeons were located in predominately middle-class neighborhoods. Singer tells us that Allen used this data to support a theory that nickelodeons "attracted a middle-class audience throughout the nickelodeon era and early teens" (Singer 9). In discussing Harlem proper, Singer ponders whether composition of a neighborhood can be used to determine the composition of that neighborhood's nickelodeon audience. He states:
While it seems reasonable to assume some degree of correspondence between the kind of neighborhood a nickelodeon inhabited and the kind of patrons it drew, it is hard to know how close such correlations actually were. One can imagine a hypothetical neighborhood, for example, whose population was half middle class and half working class but whose movie audiences were four-fifths working class. (17)
In questioning one's ability to determine the composition of an area's nickelodeon audience based on the composition of that area's population, Singer seems to be both, 1) in a general sense, providing another means of questioning Allen's argument that nickelodeons had a middle class audience, and 2) questioning Allen's characterization of Harlem proper as a middle-class neighborhood. Because Singer can only classify the socioeconomic population of Harlem proper as "mixed" (17), he questions Allen's ability to deduce nickelodeon composition based on area composition.
Despite questioning Allen's use of neighborhood composition to determine audience composition, later in the article, when Singer attempts to assess the ethnic composition of the Manhattan nickelodeon audience, he justifies use of the same method. Singer claims:
Presumably, as with the question of class composition, one could gain insight into the ethnic makeup of movie audiences (at least at neighborhood nickelodeons) by determining the ethnic makeup of the areas in which nickelodeons were located. Although neighborhood ethnic composition may not automatically have translated into movie-audience composition, it is probably fair to assume a relatively close connection in most cases. (21)
This claim that "it is probably fair to assume a relatively close connection" between neighborhood composition and movie-audience composition contradicts Singer's previous claim that "it is hard to know how close such correlations actually were." Further, the concession that ethnic composition "may not automatically have translated into movie-audience composition" cannot hide the fact that Singer is basing his argument in the next page or two on this correlation. On the following page, for example, Singer claims that "the majority of nickelodeons were in overwhelmingly Jewish areas…so one can assume that Jews constituted the largest sector of Manhattan's nickelodeon audience" (22).
Here, I believe Singer is committing the "fallacy of statistical special pleading," which Fischer defines as applying "a double standard of inference or interpretation to his evidence – one standard to evidence which sustains his generalization and another to evidence which contradicts it" (Fischer 110). In the context of Allen's argument, neighborhood composition is a questionable means of determining audience composition; in the context of Singer's argument, neighborhood composition becomes a legitimate means of determining audience composition.
I think Singer (and Allen) could also be accused of "the fallacy of statistical nonsense," which Fischer defines as "statistical mumbo jumbo" that is "in context, literally meaningless" (116). "Statistical mumbo jumbo" and "meaningless" may be a little harsh, but I do think we should question whether neighborhood composition and audience composition necessarily correlate. Singer attempts to justify his use of neighborhood composition by the fact that there is "no survey data on the ethnic composition of New York movie audiences in the Nickelodeon era" (21), but the absence of better data doesn't make the data used more meaningful.
Finally, one could argue that Singer's commitment of the "fallacy of statistical nonsense" results from his commitment of the "fallacy of counter-questions" – "an attempt at revision which becomes a mindless inversion of an earlier interpretation" (28). By his own admission, Singer wrote "Manhattan Nickelodeons" to challenge the "revisionist argument" (Allen's argument) that "stressed the importance of middle-class audiences throughout the nickelodeon era" (6). Fischer states that the result of posing a counter-question is repetition – and therefore perpetuation – of the faulty assumptions of the original. Indeed, Singer bases his conclusions on the same faulty data that Allen did.
- Dave Platt
No comments:
Post a Comment