![Sample our Language & Literature journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5941/TOC-Subject-Sample-2021-Language-Literature.jpg"})
Abstract
The problem undertaken here is to account for the relational placement in English of words traditionally known as adjectives and nouns. Two distinct orders are examined as signals of discrete meanings: one where the characterizing word is preposed to the characterized word, as in long hair, and the other where it is postposed, as in hair long. Distribution of the two signals in attested text is accounted for under the hypothesis that an Assertion of Characterization is made weaker or stronger, respectively, through this word order. With these meanings, a writer draws a distinction between Characterization the writer assumes the reader will receive as uncontested and so requires weaker Assertion and Characterization which is selected out of an array of particularly relevant possibilities and so requires stronger Assertion.
Acknowledgements
Thanks to the participants in the Seminar on Columbia School Linguistics at Columbia University for their encouragement and feedback on the research that led to this paper; Eduardo Ho-Fernández, in particular, provided useful references. The Columbia School Linguistic Society awarded graduate research fellowships to one of us (Author One) for research that led to this paper. The School of Education of The City College of New York awarded another of us (Author Two) release time for work on this project.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 For a critique of the relation between Saussure’s constructs and ours (from William Diver passim), see Davis (Citation2004: 316–324).
2 When, as is the case with the components of our hypothesized signals, there are only two items, the relationships of order and position coincide, but in general the two relationships are not exactly the same thing. For instance, the order of the overt syntactic objects would be notated differently for English finites (They give me the money) and non-finites (Giving me the money, … ): for finites, “second” and “third” in order, respectively, versus, for non-finites, “first” and “second” in order. On the other hand, the position relative to the verb could be notated identically for finites and non-finites, such as by “following immediately” and “following not immediately” (cf. Diver and Davis Citation2012: 239–245). For the purposes of this paper, order and position coincide.
3 This paper, following Columbia School practice, indicates formally hypothesized meanings in small capital letters and the names of semantic substances with initial capital letters.
4 This paper does not undertake to specify exactly what words in English can function as determiners, but analysis suggests that they include: a(n), the, any, those, my, and so forth.
5 As will become evident, we rely heavily in this paper upon written discourse, and so we often refer to “reader” and “author,” but our hypothesis appears to apply as well to listener and speaker.
6 For this system to be a viable communicative tool, it must be the case that listeners (or readers) can typically decide which word (the A) Characterizes which (the B); otherwise, there would be no way to distinguish the two orders.
7 This formulation avoids reference to traditional parts of speech (blue sky / sky blue). Cf. fn. 6.
8 An incisive critique of Halliday and Hasan (Citation1976) and of the relevance of – and irrelevance of – the study of cohesion to the analysis of linguistic elements is provided by Reid (Citation1991: 307–309). In his analysis of the grammatical signals of Number in English, Reid notes that signaled meanings of grammatical Number – the things that he posits – can indeed contribute to what he calls “textual resonance,” but for him – and for us – the message in a text is a gestalt, not analyzable into discrete fractions (e.g. the contributions of individual words).
9 “Temporal spread,” which is semantic, is actually only one of two “solutions” that Bolinger proposes; the second – reflecting the generative mindset of that day – has to do with the syntactically underlying source, through “generation,” of two types of modification.
10 One realizes too that the notion of “possible worlds” cannot but represent an enormous analytical escape hatch and abandons the testability of a hypothesis, since a language-user (such as a novelist or a liar) can imagine and represent any world he or she wishes, whether the analyst knows that world is real or not.
11 William Diver was editor of the journal WORD from 1956 to 1965. The 2012 volume was edited and published posthumously. An analysis of Diver’s work is treated in Davis (Citation2016).
12 Given language variation (even to the point of idiolects), it is a real analytical question just what language-users – and thus what data – will be admitted into one’s data set. See Davis (Citation2017b), particularly his discussion on pp. 241–242, for one way to handle that challenge.
13 One of us (Author Two) for several years taught Diver’s hypothesis at the university level, until extensive work with actual text, in collaboration with graduate students, led to a different hypothesis involving the two orders: that the order BA signals that Alternative Characterization is to be considered while the order AB signals that it is not considered. That revision, in turn, led to the present hypothesis (primarily by Author One).
14 The analysis of Diver's work that is summarized here is not published in Huffman and Davis (Citation2012).
15 In light of our definition of the signal as “[det] AB” in Diagram 1, it may help to make it explicit here that we assume the existence of meaningful null articles (one singular and one plural) in position before such examples as Ø little June (cf. a little June such as we had never seen) and Ø many little girls (cf. the many little girls who came to the party), as well as in such examples as (Ex. 3, below) Ø lead singer, Ø long hair, Ø bare chest, Ø tight pants. Such a postulation is by no means unusual, but see Huffman (Citation2001: 56) for the analysis (by Diver) that we assume here.
16 A reminder that this paper does not investigate any differences between the orders of descriptive words such as classic American summer / American classic summer, let alone long, classic American summer.
17 The count was conducted in November 2017. See footnotes 6 and 7 regarding our handling of parts of speech.
18 Counts in this paper are offered as support for our hypothesis, not as discoveries in their own right. For previous use of quantitative data in support of a signal-meaning hypothesis see Diver (Citation1987/2012: 93–95), Reid (Citation1991: 363), Huffman (Citation1997: 73, 163, 234), Reid (Citation2011: 1098–1099, 1109–1110, 1120–1126), Davis (Citation2017a: 110–114), Davis (Citation2017b: 48–49, 57, 70, 77), and Sabar (Citation2018: 21–22 et passim). This may be, however, the first use within this framework of time as an independent variable.
19 An exhaustive search in COCA yields 59 tokens of the phrase “deems the” and 1,382 tokens of “recognizes the,” most of which contain neither a BA nor an AB (these do not appear in ). The word the is included simply to focus the mechanical count on a following noun phrase.
20 See footnotes 6 and 7 as well as the section on Issues for Future Research (§6) for our handling of parts of speech.
21 Again, in this treatment, we do not treat nouns and pronouns differently, as the hypothesis appears to apply to both equally. See §6.2.2 regarding the strong tendency for the modification of pronouns to appear in the BA order (e.g. him safe).
22 An anonymous reader offers the intriguing suggestion that formal, prepared spoken discourse (as opposed to, say, spontaneous speech) may exhibit patterns of usage that resemble published written discourse. In particular, in those two communicative settings, the signal (BA) of the meaning stronger Assertion of Characterization may be more frequent than it is in, say, spontaneous speech. If so, that could well have to do with a language-user’s awareness of a relative dearth of shared situational knowledge between writer/orator and reader/listener (as opposed to casual interlocutors).
23 In general in this paper, constructed passages are used purely for illustrative purposes; we analyze only attested examples. See also §6.1.1.
24 Adjacency is also an irrelevant consideration with order (or positional) signaling in several of the previous analyses cited above, including in Huffman (Citation2002), cited extensively below. The issue of adjacency does not come up with the [det] AB signal, even in the case of strings of A’s (e.g. a long, hot summer): there all the A’s are confined between the [det] and the B, not separated from the B by a different grammatical signal.
25 Alongside the first 100 tokens of AB in Patchett (Citation2011), there were only 18 unseparated BA but 68 separated by a linking verb (e.g. was, became).
26 See §6.2.3 below regarding modification within a predicate noun (as in a wasted life).
27 In this paper, we employ the asterisk not in the generativist sense of “ungrammatical” (i.e., not generated by the grammar) but in the sense of “unattested” in our data. That is, we make no claim at all that, given human ingenuity, certain things cannot be said. For instance, imagine someone saying Gee, that was the ninthest ninth inning I’ve ever seen!
28 https://fansided.com, accessed November 2, 2016.
29 Thanks to CCNY graduate student Edward Berk for this insight.
30 Paramount Pictures. 2018. Pet sematary official trailer.
31 These words include: abed, abroad, across, adrift, afloat, afoot, ahead, alive, alone, apart, around, askew, asleep, astray, atilt, atop, awake, awash, and away.