Premise

[1.1.1] “Constructions are defined to be conventional, learned form-function pairings at varying levels of complexity and abstraction” (Goldberg 1995, 2006).

Paraphrase

[1.1.2] While syntax (defined by structures) and semantics (defined by a lexicon) were two separate domains in mainstream generative grammar, in construction grammar they are just different points on a continuum. Both are similar types of knowledge learned through usage (Heine 2011, 63). A working definition of a construction in both language and music might thus be “an entity with a conventionalized form, one that is generally paired with a particular meaning or function associated with a common situation in human communication.” In music this could mean a marked chord or progression (Hatten 1994), a conventional articulation like the half cadence, or the many schemata developed for phrases and sequences. Because music is rarely directly denotative, the notion of communicative function must be broad enough to include the evocation of mood, the suggestion of affect, and the whole range of nonverbal meanings treated in semiotics and embodied cognition.

A Linguistic Construction Exemplified

[1.2.1] Ronald Langacker, an early proponent of a construction-grammar perspective, noted that this new approach must accommodate seemingly exceptional items like idioms, even if they cannot be subsumed under general principles:

Substantial importance is given to the actual use of the linguistic system and a speaker’s knowledge of this use; the grammar is held responsible for a speaker’s knowledge of the full range of linguistic conventions, regardless of whether these conventions can be subsumed under more general statements. [The construction model is] a non-reductive approach to linguistic structure that employs fully articulated schematic networks and emphasizes the importance of low-level schemas. (Langacker 1987, 494)

An idiomatic construction now famous among construction grammarians may help to illustrate what Langacker had in mind. The old comedy line “Waiter! What’s this fly doing in my soup?” is an instance of the “What’s X doing Y?” construction (or WXDY? for short), a common pattern studied extensively by Paul Kay and Charles Fillmore (Kay and Fillmore 1999). Note that the contextual meaning of the WXDY? construction is associated with surprise due to the incongruity of a scene, and usually accompanied by speaker disapproval (Kay and Fillmore 1999; Bybee 2010, 28). The speaker is not merely asking what X is doing. Instead, the speaker is challenging the larger picture or the motivation behind the action. That extra meaning, something beyond what could be looked up in a dictionary, is learned as part of the WXDY? construction. Constructions encapsulate rich experiences as rich memories, and construction grammars assume a far greater engagement with memory than did mainstream generative grammars.

[1.2.2] Among native speakers of English, the WXDY? construction is so well remembered and understood that, with all its features and inferences intact, it made a reliably comic appearance in the film Ghostbusters (1984):

Dana Barrett: [reading from the printout] “Zuul was the minion of Gozer.” What’s Gozer?
Dr. Peter Venkman: Gozer was very big in Sumeria.
Dana Barrett: Well, what’s he doing in my ice box? [=WXDY?]
Dr. Peter Venkman: I’m working on that.

[1.2.3] In the paper by James and colleagues cited earlier (James et al. 2014, 1), the authors speculate that “the processing of syntax and meaning may coincide in complex intra-musical contexts.” This is consistent with a view of a construction as the merging of a structure with a meaning. The linguists Johnson and Goldberg (2013) conducted a number of experiments testing construction grammar using Jabberwocky sentences (grammatically correct passages that use nonsense words, as in the Jabberwocky poem in Alice in Wonderland: “Twas brillig, and the slithy toves. . . .”). They found that “phrasal abstract constructions are associated with semantics [e.g., meanings] even when they contain no open-class lexical items [e.g., standard nouns], and that the meaning is accessed quickly and without explicit instruction” (2013, 1451). Although nonsense words were used, listeners were able to infer (subjective) meaning from the constructions. This is further evidence that constructions are learned as form-meaning pairings, and may suggest how musical meaning may still be conveyed in the absence of obvious denotation.

A Musical Schema Viewed as a Construction

[1.3.1] For the touring virtuoso composer-pianists of the late nineteenth century, the arena for human communication was on stage in live performance. The show business of the virtuoso required managing the mood and expectations of the audience, and successful constructions developed by one artist to amaze or guide an audience could be quickly adopted by other artists. The fifteen-year-old Rachmaninoff, for instance, learned a great deal from his apprenticeship with Anton Arensky at the Moscow Conservatory. Arensky was a virtuoso pianist who pioneered the composition of bravura works for two pianists, notably a suite for two pianos titled Silhouettes (op. 23, 1892). Rachmaninoff would go on to publish a series of works for two pianists later in the same decade, so he had a practical interest in learning successful constructions used by Arensky.

[1.3.2] Patrick McCreless (2006) has described a general pattern in many nineteenth- and twentieth-century compositions where near the conclusion there is often a big descent and “crash” before the final denouement. In the last nine measures of Silhouettes (see Example 1), Arensky asks his duo pianists to perform a very particular case of that general pattern. They play a brilliant cascade descending across the whole keyboard before ending with the conventional “Final Fall” (the high-to-low octaves on the final C-major chords; see Gjerdingen 2007a, 168). Harmonically, Arensky’s rapid cascade involves the tones F♯ and A acting as appoggiaturas to G within the persistent sound of a C-major tonic triad. Rhythmically, Arensky sets up a series of repeating five-eighth-note groups at odds with the duple meter.

[1.3.3] After the Russian Revolution, Rachmaninoff began his long second career as a traveling virtuoso pianist in the West. On several occasions he toured as a duo with the violinist Fritz Kreisler, and he no doubt had to serve as accompanist for Kreisler’s well-known encore pieces Liebesleid and Liebesfreud (“Love’s Sorrow” and “Love’s Joy”). He eventually wrote virtuoso solo-piano fantasies on both encores. In the final nine measures of the Liebesleid fantasy (1921) he faced a situation very much like that at the end of Arensky’s Silhouettes. Fully in keeping with the maxims of construction grammarians, Rachmaninoff chose Arensky’s cascade and Final Fall as a prêt-à-porter item “off the rack,” something already fitted for a stock situation by his old master (see Example 2 in comparison with Example 1).

[1.3.4] He retained Arensky’s harmonic color, the descending cascade, the virtuosic speed, the rhythmic play (a hemiola beginning in m. 215), and the Final Fall—in short, the whole cluster of features that characterized Arensky’s model construction. And of course the contextual meaning of the passage, the performative message to the audience to “be amazed and get ready to applaud,” is fused with the technical structure of the passage. Nowhere in a dictionary definition of “chromatic neighbor tones” would it say “be amazed and get ready to applaud.”

[1.3.5] Rachmaninoff chose to traverse the gap between his lowest note (m. 217) and the first chord of the Final Fall with an arpeggio (mm. 217–18), whereas Arensky’s pianists simply lifted their arms across the same space. We view this as a minor difference within the range of options for this construction, and less of a real difference from an embodied perspective. The developmental psychologist Michael Tomasello has defined a construction as “a unit of language that comprises multiple linguistic elements used together for a relatively coherent communicative function” (Tomasello 2003, 100). Were we to replace his words “language” and “linguistic” with “music” and “musical,” his definition would reflect our view of musical constructions. Some types of constructions are replicated literally, almost note for note; others have variable “slots” that permit a limited range of alternatives.

[1.3.6] Given the same communicative opportunity, Rachmaninoff often chose this same construction. Ten measures from the end of his Second Piano Concerto (1900–1901), for instance, he launches into nearly the same cascade, with the first bar shown in Example 3 being almost a transcription of the close of Arensky’s Silhouettes (cf. Example 1; the correspondence is even closer in relation to Arensky’s unreduced original).

[1.3.7] What looks like an F♯ half-diminished chord in Examples 1, 2, and the first bar of Example 3—what a harmony book would likely describe as a “nonfunctional” or “coloristic” chord—is thoroughly functional in this frequently replicated construction. For Arensky and Rachmaninoff, “function” was determined not by some universal principle or a rule-based theory of tonal syntax but by how a listener interpreted the interplay of sonorities within what we will term this “bringing down the house” construction. As Tomasello writes of language acquisition, “The notions of communicative intention and function are correlative. Someone uses a piece of language with a certain communicative intention, and so we may say that that piece of language has a certain function” (Tomasello 2003, 3). The communicative intent of the “bringing down the house” construction was obvious to experienced concertgoers, and thus its component sonorities accrued functionality from listeners’ abilities to recognize that intent in the context of a bravura finale.

[1.3.8] Listeners experiencing the blindingly fast close of Mily Balakirev’s showpiece Islamey (1869; see Example 4) might hear a few more chromatic neighbor tones than in the similar passages by Rachmaninoff or Arensky, but they would recognize the same communicative intention, the same physicality, the same glistening major sonority, and all the collocated features that make this construction so memorable for a listener and made it so effective for the composer-pianist.

[1.3.9] In language, the shared knowledge of constructions enables many types of wordplay. A wry musicologist, one knowing the construction “That’s so [DATE]”—with its associated meaning of “that item is passé, as out of date as [name or date of a past era]” (Wee and Tan 2008)—might hear a Bach ricercare and say to a colleague, “That’s so 1690s.” For those who share the requisite frames of reference, construction-based wordplay celebrates knowledge held in common and reinforces cultural bonds. Thus a concert-goer in the 1890s, one who shared an implicit understanding of the “bringing down the house” construction, might have smiled knowingly and approvingly when the young Glazunov (trained from an early age by Rimsky-Korsakov) concluded his Etude (op. 31, no. 1, see Example 5) with that fashionable construction presented upside down. This was playful tinkering (Meyer 1980), not a Russian revolution— Glazunov would go on to become director of the St. Petersburg Conservatory.