Page 65 - Summer2017
P. 65

  categories with various distances from one another was popularized, most notably by Shepard (1957), and has been utilized in a large body of subsequent research. This literature has afforded the mathemati- cal operations of multidimensional scaling (MDS), developed by Togerson (1952) and others. MDS takes the measures of association between individual cat- egories and distills them into a low-dimensional rep- resentation of the space inhabited by the categories, with their relative locations within this lower dimen- sion space (Kruskal and Wish, 1978). (See Hout et al. [2013] for an overview of this technique and its ap- plications.)
Figure 1. Results of the multidimensional scaling (MDS) analysis on mem- bers’ primary and secondary interests in the 13 technical areas of the Acous- tical Society of America. Data are from membership records from Fall 2014.
 This was exactly what we needed to situate the techni-
cal areas in the conference venue, projecting the relation- ships between the 13 technical areas onto the (at most) 3 dimensions of the hotel space.
As input, we took ASA membership information concern- ing the primary and secondary areas of interest of 7,126 members, summed up in Table 1. Members choosing pri- mary and secondary areas of interest indicated a proximity of the two areas.
The counts were submitted to an MDS analysis using the ALSCAL scaling algorithm (Takane et al., 1977) in SPSS. ALSCAL calculates the optimal distances between catego- ries in an n-dimensional stimulus space by minimizing a stress value (“Kruskal’s S-Stress”). The stress value is used to choose the optimal number of dimensions where any added dimension does not significantly improve the fit.
For our analysis, this process conveniently yielded two di- mensions, shown in Figure 1. In the resulting target-shaped space, similar technical areas are located close to one another and dissimilar technical areas are further apart. Around the perimeter is a circular progression running (starting at bot- tom left) from Speech Communication to Psychological and Physiological Acoustics to Musical Acoustics to Architectural Acoustics and Noise to Structural Vibration and Engineering Acoustics to Physical Acoustics to Underwater Acoustics to Acoustical Oceanography to Biomedical Acoustics to Animal Bioacoustics and then back to Speech Communication.
The orientation of this perimeter yields two dimensions, each with a general interpretation. The vertical dimension corresponds to the distinction between a disciplinary focus on artificial and engineered entities at the top and organic entities in the wild at the bottom. On the horizontal axis, the
areas vary from human and cultural to the left (e.g., music and architecture) and nonhuman (e.g., oceans) to the right.
In terms of how the ASA itself coheres, the MDS analysis suggests an answer. All disciplines are connected by a chain of disciplines made up of pairs of neighbors, each of which have very obvious and important shared interests. They are further connected by Signal Processing, which sits in the cen- ter and apparently shares members with all other technical areas.
Armed with these results, the sequence of technical neigh- bors was laid out on the two floors of the Indianapolis Mar- riott, roughly as in the top and bottom halves of the MDS plot in Figure 1. More generally, what our analysis shows is how such a diverse organization hangs together. It is a net- work of allied pairs or triplets of disciplines, each overlap- ping with different neighbors to form a coherent whole, with Signal Processing as a universal interest.
Biosketches
Kenneth de Jong is professor of linguis- tics at Indiana University, conducting research into the diversity of processes pertaining to speech production and perception, from motor coordination to the acoustic structure of linguistic categories to the acquisition of percep-
tual abilities in a second language. He was editor in chief of the Journal of Phonetics, is president of the Association for Laboratory Phonology, and, most pertinent to this article, was cochair of the 168th meeting of the Acoustical Society of America in Indianapolis.
   Summer 2017 | Acoustics Today | 63




















































































   63   64   65   66   67