About CA
Contact CA

Search Journal

Editorial Board

Permission to Reprint


Site Map


Autographic and Allographic Imitation: Revisiting Counterfeit in Linguistic and Musical Arts

  Jeremy Orosz

Imitation within sonic arts, linguistic, musical, or otherwise, broadly defined, is a heterogeneous set of practices ranging from comical impersonation to outright counterfeit or forgery. This paper provides a taxonomy of imitation and mimicry within both language and music, dividing each into two respective categories, autographic imitation and allographic imitation, the terms for which are repurposed from Nelson Goodman’s Languages of Art.

Key Words
autographic; allographic; forgery; Nelson Goodman; imitation; language; music; substitution


1.  Introduction

Nelson Goodman’s Languages of Art (1968) is a seminal work in aesthetics that continues to inspire commentary from authors in a wide range of fields. Perhaps its central theoretical tenet is Goodman’s proposal of a distinction between autographic and allographic arts. Painting and sculpture are autographic arts, in which there is one definitive object that comprises a work. Music, like literature and drama, is an allographic art that allows infinite, equally viable realizations of the same work, perhaps allographs (my term) of the same grapheme (again, my term) reproduced through a notational system.[1] Though this classificatory scheme remains a useful heuristic, Jerrold Levinson, Gerard Genette, and others have convincingly argued that the autographic-allographic distinction is not as neat as Goodman suggests. Music and literature, for example, do not strictly reside within the allographic domain.[2] Even if Goodman’s categories do not adequately distinguish the arts from one another, rather than discard this binary out of hand, I will argue that the terms ‘autographic’ and ‘allographic’ are of great utility for classifying practices of counterfeit and mimicry within individual arts. In particular, the acts of reproduction, forgery, and imitation, in both linguistic and musical arts, most clearly demonstrate the simultaneous autography and allography within the respective systems of language and music.

2. Autographic and allographic arts

The essential difference between autographic and allographic arts, according to Goodman, is the possibility of forgery. A work of art is considered autographic “if and only if the distinction between original and forgery… is significant.”[3] In Goodman’s words:

[I]n music, unlike painting, there is no such thing as a forgery of a known work. There are indeed, compositions falsely purporting to be by Haydn as there are paintings falsely purporting to be by Rembrandt; but of the London Symphony [sic], unlike the Lucretia, there can be no forgeries. Haydn’s manuscript is no more genuine an instance of the score than is a printed copy off the press this morning.[4]

Although forgery is not practiced within the respective notational systems of music and language, Goodman was misguided in dismissing the possibility of forgery in the arts he classified as allographic. Sound is forgeable, even if notation is not.[5] True, one cannot forge a score for one of Haydn’s London Symphonies but it is possible to create a counterfeit version of the Clash’s London Calling album. Replicating a recorded sonic document is much akin to creating a copy of Botticelli’s The Birth of Venus, in that a vocalist mimicking the precise timbre of a famous singer and a painter emulating every brushstroke of an iconic painting face analogous challenges. If we recognize this as such, according to Goodman’s own definitions recorded sound belongs among the autographic arts, in which every feature is constitutive of a work and “no deviation is insignificant.”[6] Theodore Gracyk makes a similar argument, suggesting “that precise details of timbre and articulation can be essential properties of a musical work” that renders recorded music as “autographic, because notational determination is entirely irrelevant to the genuineness of its instantiations.”[7] Only music that is transmitted primarily through a notated score, say, Western art music of the Common Practice Period, should be considered strictly allographic. 

The sections that follow will respectively explore autographic and allographic imitation within the sonic arts. Emulation of a particular sonic source or of a specific sonic performance is a form of imitation that is autographic in nature, while stylistic mimicry of the writing or compositional style of a particular artist is allographic. The forms of autographic imitation in Table 1 involve replication of sound through sound, while the types of allographic imitation are instead transmitted through notational systems for reproducing sound.[8]

Table 1:  Imitation in Music and Language

3.  Autographic imitation of language

Autographic imitation within both language and music is divisible into three practices:  replication, imitative substitution, and whimsical impersonation.The boundaries between these subcategories are, to a degree, porous, as we shall see below, but the differences between them are clear enough to merit distinguishing each from the others.

Autographic imitation of linguistic sound includes all attempts 1) to mimic an imitated person with the utmost accuracy, allowing the possibility of deception (replicative), 2) to provide a passable replacement for a speaker in their absence (substitutive), or 3) to perform an impersonation to humorous ends (whimsical). 

True replication of another person’s voice, in which a vocal chameleon could convince a listener that their voice is that of another person, although rare outside of fictional narratives is plenty common on stage and on screen. An audience might suspend their disbelief that the characters within a fictional world could be fooled by an imposter’s voice. A classic case of this is found in Edmund Rostand’s drama, Cyrano de Bergerac, in which a handsome but simple soldier, Christian, strives to win the affections of the much-desired Roxanne. In the iconic balcony scene, the poetically precocious Cyrano stands in the shadows, feeding Christian his lines, helping him to more elegantly profess his love for Roxanne. When this proves cumbersome, Cyrano pulls the plug on the operation and simply emulates Christian’s voice, delivering impeccable lines of improvised verse while hiding in the garden below. When Roxanne notes that the speaker’s “words have learned to climb” to reach her at a more natural pace, Cyrano explains this away suggesting that “with practice, such gymnastic grows less hard.” When she remarks that his “voice rings with a tone that’s new,” he declares, “I dare to be myself for once—at last!” Roxanne apparently deems this explanation satisfactory and remains blissfully unaware that the fair soldier of her dreams is barely literate.[9]

Outside of fiction, using the voice of one person as a substitute for another is more common than is replication, deceptive or otherwise. Perhaps the clearest case of imitative substitution of spoken language is found in the recasting of voice actors. In some cases, the new actor’s voice might actually pass as that of their predecessor. Watching the 2011 Muppets reboot, some viewers might fail to notice that Jim Henson and Frank Oz no longer provide the voices for the familiar puppets. More typically, however, voice actors are not held to the standard of being able to convince an audience that no recasting has occurred. 

When a live-action film or television franchise is reformatted as a cartoon, the original screen actors rarely speak for to their animated avatars.[10] Both the television cartoon The Real Ghostbusters (1986-1991), based upon the wildly successful film Ghostbusters (1984), and the Star Wars prequel spin-off The Clone Wars (2008-2014) hired substitutes to replace the voices of the celebrities who starred in the big-screen blockbusters.[11] Although few would mistake the performance of the actors on either show with the voice of their more prominent counterparts, all characters are voiced by a more than capable substitute.[12] Audiences seem unperturbed by minor differences in voice when the likeness of a film actor is rendered in cartoon form; a substitute actor that matches the tessitura and accent of the person imitated is generally deemed sufficient.[13]

The above examples of counterfeit of spoken language all have a shared practical goal of passing for or replacing another person, yet mimicry of another’s voice can, of course, be done merely for whimsical or humorous ends, often with the explicit goal of parody, or even caricature, if sufficiently exaggerated. Celebrity impersonation has long been the bread and butter of comedians, whether performing a stand-up routine or in a sketch on a program, like Saturday Night Live or Key and Peele

4.  Autographic imitation of music

The category of autographic imitation of music includes all attempts to reproduce either a specific performance, most likely captured on a recording, or the performative style of a known musician or ensemble 1) to mimic the imitated performer or ensemble with the utmost accuracy, allowing the possibility of deception (replicative), 2) to provide a passable replacement for a performer in their absence (imitative substitution), or 3) to mimic another musician or group, to humorous ends (whimsical impersonation). 

Replicative forgery of musical recordings is ubiquitous; studios around the world produce counterfeit recordings of popular tunes. It remains more economically advantageous to hire musicians and recording technicians to produce sound-alike versions than it is to pay the rights to distribute the authentic recordings. In fact, one such group that specialized in creating knockoffs called themselves, quite cleverly, “The Original Artists,” so that they could say, in earnest, if not wholly in truth, that “all songs are recorded by ‘The Original Artists.’”[14]

Just as deceptive replication of a speaking voice is common within fictional narratives, the same phenomenon occurs with non-diegetic singing in musical drama. In Franco Alfano’s opera version of Cyrano de Bergerac, the aforementioned balcony scene is performed by singing rather than speaking. Likewise, the plot of Puccini’s Gianni Schicchi revolves around the title character convincing a notary and two witnesses that his voice is, in fact, that of another: a literal case of forgery. Another case of deceptive imitation within fiction is found in the recent film The Death of Stalin (2017), in which the dictator demands a recording of a performance of Mozart’s Piano Concerto No. 23 (K. 488). The concert had not been recorded, however, and a comedy of errors ensues as the engineers scramble to replicate every detail of the initial performance, including audience applause, to produce a counterfeit record for their esteemed comrade.[15]

Substitution of one musician for another is impossibly common, but rarely is this a case of imitation. If an orchestra hires a new concertmaster, the new player does not emulate their predecessor in any non-trivial way. Imitative substitution comes into play most often when a singer with a distinctive voice needs to be replaced. When vocalist Rob Halford departed from the band Judas Priest, a singer in a Judas Priest tribute band, already an experienced imitator, was recruited to take his place. Replacement of some original cast members in popular Broadway shows, say, Lin Manuel Miranda in his brainchild, Hamilton, might also require imitative substitution, but only in cases where a singer’s unique talents are difficult to reproduce.   

Just as impressionists served as the primary example of whimsical imitation of spoken language, some musicians make their career as mockingbirds through song, mimicking the performative style of well-known vocalists. New York-based singer and actress Christina Bianco became a minor internet sensation for performing a lengthy rendition of “Total Eclipse of the Heart” in the vocal stylings of at least a dozen famous divas, from Julie Andrews to Christina Aguilera. Comical imitation of instrumentalists, though not nearly as common as that of vocalists, is likewise possible, if the performer is known for his or her singular idiolect. Glenn Gould’s idiosyncratic performances have long been a target for parody; pianists delight in lampooning the late virtuoso, in platforms ranging from high-profile films to amateur YouTube videos.[16] As with whimsical autographic imitation of language, sufficient exaggeration of a performer’s style results in caricature.  

5 . Allographic imitation of language

Allographic imitation of sound, like its autographic counterpart, is roughly divisible into the same three categories of replicative, substitutive, and whimsical imitation. Allographic imitative practices demand further subdivision, as shown in Table 2 below.

Table 2:  Categories of Allographic Imitation[17]

Allographic imitation of language, a literary phenomenon, includes any attempt to reproduce the writing style of another author, either 1) to mimic the style of a writer with the utmost accuracy, allowing the possibility of deception (replicative), 2) to replace an author in their absence (substitutive), or 3) to more generally produce a pastiche or parody of an author’s work or a style (whimsical).

Replication of written language encompasses a diverse range of practices in which a work is ascribed to someone other than its true author. Artists, regardless of medium, have been known to sell their work under the name of a more famous counterpart to demand a higher price.  Perhaps more common in literature is presenting one’s work as that of a fictional author. Three British authors did exactly this in the eighteenth-century. Horace Walpole presented his 1764 novel, The Castle of Otranto, as a medieval relic. Thomas Chatterton did the same with poetry, falsely ascribing his work to fictional medieval poet Thomas Rowley. Perhaps the most successful hoax, James MacPherson published a collection of poems in the 1760s said to have been written in the Dark Ages by a poet named Ossian. These poems attained great popularity—Napoleon was fond of them and even demanded that paintings be made featuring the fabled Ossian—before they were revealed half a century later to be the work of the late Scotsman.   

Another type of replicative, and perhaps deceptive, imitation is when an author’s work is uncredited, as with a ghostwriter of a biography, or a secretly multi-authored work published under a single name. Authors of young adult fiction, who publish short books at a blistering pace, must surely have teams of assistants who write collectively under a single pen name.  

Imitative substitution is likewise a common phenomenon in the literary arts. Television series often employ a small army of writers; even episodes from the same season of a show may have different authors. Movie sequels typically have scripts produced by a new team, as do, unfortunately, musical scores (more on this below). Although far less common than in television and film, series of novels are occasionally completed by other authors, often after the original author has died.[18] All of these practices are necessarily imitative, as a new author in a television, film, or novel series must ensure that the characters always "act like themselves," so to speak, neither stepping out of character nor devolving into caricatures thereof, continuing the established norms of what a fictional character would do and say. Unauthorized, non-canonical sequels, ranging from fan fiction to more lofty attempts at continuing beloved texts, such as Alexandra Ripley’s Scarlett, conceived as sequel to Gone with the Wind, perhaps straddle the line between substitutive and whimsical imitation.   

Literary imitation without the objective of replication or substitution is endemic to (post)modernity. The works of James Joyce are known for their inclusion of light-hearted imitations, parodic or otherwise, of the literary idiolects of countless authors, and writing “in the style of” another remains popular to this day. It is not always practical to determine whether or not a whimsical imitation of an author is meant to parody, but when the imitation drifts from uncanny similarity to exaggerated lampoon, this may be deemed caricature. (Table 2, above, illustrates that there is no clear dividing-line between pastiche and parody.)

6.  Allographic imitation of music

Allographic imitation of music includes any attempt to reproduce the style of another composer with the goal of 1) mimicking a composer’s voice with the utmost accuracy, allowing the possibility of deception (replicative), 2) providing a proxy for the work of another composer in their absence (substitutive), or 3) producing a pastiche or parody of a composer’s work (whimsical). 

Truly forged compositions presented as the work of another, though rare, do surface from time to time. Perhaps the clearest example, in relatively recent memory, is the case of Winifred Michel’s fabricated discovery of six lost sonatas of Joseph Haydn. Michel’s forgeries were so cleverly constructed that he managed to convince even the most preeminent Haydn scholars that his spuriously sourced documents were, in fact, the early works of the Austrian master.[19] It was not the musical content of the forgeries that led to their discovery as counterfeit but rather the manner in which they were produced. Michel was found out because the pen used to forge the score would not have been available to Haydn.[20] The lesson here seems to be, as with making a counterfeit bank note, that one must take care to use the proper ink.

As with literature, another form of replicative allographic imitation is uncredited ghost authorship. Defining the limits of musical authorship is, of course, a tricky matter. Is a drummer in a band a co-composer for designing his or her own fills? Is a baroque keyboardist likewise an author for deciding how to realize the harmonies prescribed by figured-bass symbols? Questions of this nature, though worthy of answers, will be put on ice at present with the use of an admittedly old-fashioned, Eurocentric definition of composition, restricting it to the act of creating a score in which all notes and rhythms (at a minimum) are specified.[21] Within this intentionally restrictive, specifically modern definition, few proven cases of uncredited musical authorship come to mind. Contested or unknown authorship, however, is plenty common. After the death of Italian composer Giacinto Scelsi, his sometime collaborator Vieri Tosatti claimed to have been the true composer of a portion of Scelsi’s work, a claim yet to verified or refuted.[22]

Posthumous completion of a musical work is perhaps the quintessential case of allographic imitative substitution. Most typically, surrogate composers explicitly adopt the language of their fallen predecessor. Alban Berg died without having orchestrated the third act of Lulu. Fredrich Cerha provided the orchestral rendering more than forty years after Berg’s death, emulating Berg’s scoring habits to the best of his own ability. In a similar story, Puccini succumbed to throat cancer before completing Turandot, leaving only sketches for much of the third act. Franco Alfano, mentioned above for his Cyrano opera, initially brought the work to life; years later, the late Luciano Berio produced his own completion of Act III. With such explicit counterfeits, there is perhaps room for more than one authentic version. If two competing renderings of a posthumously completed work each provide a reasonable imitation of the composer’s style, selecting one of the two as the authoritative version would rely on arbitrary criteria. For example, do we prefer the version that leans on the composer’s characteristic figures more heavily, risking caricature through exaggeration, or the version that treads more lightly, imitating the composer less conspicuously?                                                                  

While the realization of the works discussed above required imitation, it should be made clear that not all posthumous completion is necessarily imitative. The storied case of the tenth symphony, an apparently common cause of death for (Germanic) composers, may serve to clarify this distinction. When Mahler left behind the torso of his tenth symphony, Deryck Cooke aimed to make it performable by replicating Mahler’s late style to fill in the gaps between the completed passages.[23] Brian Newbould’s posthumous completion of Schubert’s Symphony No. 10 is an almost identical case. Berio’s Rendering, however, which incorporates fragments of Schubert’s tenth symphony into a patchwork of music in Berio’s own style, is clearly non-imitative, as there is no attempt to emulate Schubert’s compositional voice. One might argue that certain movements of Süssmayr’s completion of Mozart’s requiem are also non-imitative; if Süssmayr did intend to mimic Mozart’s style, he was clearly, at times, unsuccessful.

As with hiring new screen writers, parachuting in a new composer into a television or film series can also be a form of imitative substitution. John Williams composed the music for the first installments of the Superman, Harry Potter, Star Wars, and Jurassic Park franchises but left other composers to continue these respective series. Michael Giacchino has, of late, become the heir apparent to Williams, landing jobs scoring recent blockbusters such as Jurassic World (2015) and Rogue One: A Star Wars Story (2016).[24] In the music for these films, Giacchino takes themes originally composed by Williams and seamlessly weaves them into his scores, which are stylistically analogous to the hegemonic Hollywood style for which Williams is known.[25]

Whimsical imitation of musical styles has been practiced at least since the eighteenth century. For the last three centuries, composers have used styles alien to their typical idiolect, with or without irony, in the form of pastiche. From Mozart's radical juxtapositions of earlier eighteenth-century styles, well documented in the discourse surrounding “The Musical Topic,” to contemporary artists like Lady Gaga and Bruno Mars, whose retro chart-topping hits unabashedly recall the music of the 1970s, 1980s, and 1990s, pastiche remains ubiquitous.[26] Allographic caricature of music is likewise possible, if a composer imitates the style of another in an appropriately exaggerated manner. 

7.  Concluding thoughts

This paper provides a taxonomy of imitation in linguistic and musical arts, demonstrating that Goodman’s immediate dismissal of autography within musical and literary arts is unfounded. We have seen that the categories of autographic and allographic imitation, and their respective subcategories of replication, imitative substitution, and whimsical imitation, classify a broad range of practices. Some examples, as argued above, ride the line between sub-categories, yet this taxonomic scheme nevertheless provides what I believe to be a clear map of the terrain of imitation within the sonic arts.

Jeremy Orosz

Jeremy Orosz is an Assistant Professor of Music at the University of Memphis.  He earned both his M.A. and Ph.D. in Music Theory at the University of Minnesota, where he also pursued a master's degree in linguistics.  He has published academic articles on a variety of topics at the intersections of both disciplines, and has presented papers at music theory, musicology, and other interdisciplinary venues across North and South America. 

Published on June, 14, 2018.



*Thanks are due to the anonymous reviewers for their insightful comments on an earlier draft of this article.

[1] Goodman also classifies architecture as an allographic art because, he argues, “any building that conforms to the plans and specifications… is as original an instance of the work as any other.” Nelson Goodman, Languages of Art (Indianapolis: Hackett, 1976/1968), p. 120.

[2] The autographic/allographic binary been subject to substantial scrutiny over the past half century. One of the highest profile critiques of Goodman’s terms is found in Levinson’s “Autographic and Allographic Art Revisited,” in which the author suggests that Goodman is guilty of logical fallacy. See Jerrold Levinson, “Autographic and Allographic Art Revisited,” Philosophical Studies 38 (1980), 367-83. Gerard Genette offers a concurring review, devoting an entire book to revising Goodman’s ideas. See Gerard Genette, The Work of Art:  Immanence and Transcendence, trans. G.M. Goshgarian (Ithaca and London: Cornell University Press, 1997).

[3] Goodman, Languages of Art, p. 113

[4] Ibid., p. 112. Gerard Genette offers a similar analysis of forgery in different arts, arguing that with “painting, the production of fakes or forgeries… is really an existing practice.” Rather, with “literature and music, forgery is not practiced, because a correct copy of a text or score is simply a new copy…neither more nor less valid from a literary or musical point of view, than the original.” Ibid., p. 15-16.

[5] Goodman acknowledges the secondary nature of notation to musical composition, calling a score “dispensable,” noting that “music can be composed and learned and played ‘by ear’ … even by people who cannot read and write any notation.” See Goodman, Languages of Art ,127.

[6] Ibid., p. 116

[7] Theodore Gracyk, Rhythm and Noise: An Aesthetics of Rock (Durham: Duke University Press, 1996), p. 32.

[8] The realization of an allographic (compositional) musical imitation, however, may require autographic imitation of an appropriate performative style to produce a convincing counterfeit. 

[9] Such vocal forgery is likewise common in television and film. Tina Fey’s comedy series 30 Rock features several episodes that revolve around impressionists and their deceptive tricks.  Impersonation is also a common trope in science fiction, in which a character uses a form of technology that alters their voice to aid with espionage or subterfuge.

[10] The Star Trek original cast returned to voice their characters on the 1970s cartoon spinoff, but this is perhaps the exception rather than the norm. See http://www.imdb.com/title/tt0069637/, accessed August 2017.

[11] See http://www.imdb.com/title/tt0458290/, accessed August 2017.

[12] The possible exception is Corey Burton’s remarkable mimicry of Christopher Lee as Star Wars villain Count Dooku.

[13] The same is not true, however, of characters without a material human form. The actors who voiced the mechanical and spiritual beings in the Star Wars films were indispensable to the Clone Wars cartoon, as these characters are defined as much by their voice as their likeness. Most prominently, Dee Bradley Baker continued to play the cyborg C-3PO in the cartoon, and, perhaps surprisingly, Liam Neeson returned to give voice to the spirit of Qui-Gon Jinn.  

[14] I am grateful to Thomas Robinson for bringing this to my attention.

[15] I thank the anonymous reviewer for making me aware of this example.

[16] See Julie Brown, “Channeling Glenn Gould: Masculinities in Television and New Hollywood,” in James Deaville (ed.) Music in Television: Channels of Listening, (New York:  Routledge, 2011).

[17] Posthumous completion of a single, unfinished work, discussed at in section 6, is much more common with music than with literature, whereas unauthorized sequels, like fan fiction, though a ubiquitous literary phenomenon, form an empty category in music, except perhaps for tribute bands.

[18] For example, the late Stieg Larsson, a Swedish author known for his millennium trilogy of novels and their subsequent film adaptations (“The Girl With…” books) planned to write several more installments to continue the series. David Lagercrantz has since taken over, despite the protestations of Larsson’s surviving partner, publishing The Girl in the Spider’s Web in 2015 and The Girl who takes an Eye for an Eye in 2017. (For discussion of this, see https://www.theguardian.com/books/2015/mar/31/stieg-larsson-millennium-novels-sequel-the-girl-in-the-spiders-web, accessed August 2017).

[19] Allan Konzinn, “Discovered Sonatas May Be Faked Haydn,” The New York Times (Dec 28, 1993), C17.

[20] Michael Beckerman. “All Right, So Maybe Haydn Didn’t Write Them. So What?:”  CLASSICAL VIEW, The New York Times (May 15, 1994), H33.

[21] Such a notion of authorship is necessarily incompatible with jazz, in which the line between improvisation and composition is fuzzy. Tellingly, jazz performers often refer to musical notation documents as charts rather than as scores; the difference in terminology betrays the more limited role of notation in jazz practice.  

[22] “Composer and longtime Scelsi collaborator Vieri Tosatti published a blistering article in Il Giornale della Musica, claiming outright authorship of Scelsi’s works. The ensuing controversy reached its apogee in the Piano Time’s March 1989 issue, in which musicologist Guido Zaccagnini (b. 1952) invited a group of noted composers, writers, and arts operators to discuss the matter. The group met and “engaged in a lively discussion that… did not settle the Scelsi-Tosatti controversy.” See Franco Sciannameo and Alessandra Carlotta Pellegrini, “Introduction,” in Music as Dream: essays on Giacinto Scelsi (Plymouth: Scarecrow, 2013), xi. Further discussion can be found in Guido Zaccagnini, “Is Composing a Cooperative Effort? A Historical Roundtable,” in Franco Sciannameo and Alessandra Carlotta Pellegrini (eds.), Music as Dream, pp. 1-10.

[23] Only the first movement was essentially complete; the others were drafted in short score, at most. Consequently, this work is sometimes listed as Mahler/Cooke in concert programs and recording credits.

[24] Of note, Giacchino launched his career by imitating Williams, composing music for The Lost World video game, adapted from the film for which Williams had composed the score. A complete overview of his work can be found on his personal website:  http://www.michaelgiacchinomusic.com/, accessed August 2017.

[25] Other composers who have taken up the baton of a film franchise after Williams’s departure shadow his style less closely. Patrick Doyle, who replaced Williams for the fourth installment of the Harry Potter series, despite retaining some of the original themes produced a score that sounds little like Williams’s work.

[26] The field of topic theory stems from Leonard Ratner, Classic Music:  Expression, Form, and Style (New York:  Schirmer, 1980).  A summary of the advancements in this area can be found in Danuta Mirka (ed.), The Oxford Handbook of Topic Theory (New York:  Oxford, 2014).