Quotation as an interface phenomenon

1 Quotation types and interpretational flexibility

Quotation is a metalinguistic device used to talk about certain dimensions of language (see, e.g., Cappelen and Lepore 1997; Davidson 1979; Saka 1998). In quotational constructions, expressions are mentioned rather than or in addition to being used denotationally. With an assertion like in (1a), for example, in contrast to (1b), the syllabic setup of the word sofa is described and the quotation marks (or quotes for short) around sofa indicate this use, which means reference is made to a linguistic dimension of the quoted expression (see, e.g., Quine 1981).

(1a) is an example of pure quotation. Besides pure quotation, quotes are used, among other things, to signal scare quotation, as in (2a), direct quotation as in (2b) as well as mixed quotation as in (2c) (see, e.g., Brendel et al. 2011; Saka and Johnson 2017).

Quotation is a type of communicative act in which meaningfulness is not achieved through the pairing of a form with a sense but through the demonstration of a form. By mentioning some utterance or some word or group of words, quotations display an expression but do not describe it (Clark and Gerrig 1990). A common definition of their semantics holds that quotations refer to the expression contained within the quotes reflexively (see Ludwig and Ray 2017: 102). The referential difference between a denotationally used expression and a mentioned expression occurring in quotations as in (1a) is reflected, for instance, in the incompatibility of the mentioned noun with a determiner, as is illustrated in (3).

The issue of quotation and the use of quotes has a long history in language philosophy. We refer the reader to Cappelen et al. (2019) for an overview of the different theories (e.g., Demonstrative Theory, Description Theory, Identity Theory). An orthodox definition holds that quotes are used to refer to a linguistic shape by pointing to something that has this shape. This is the gist of Davidson’s Demonstrative Theory (Davidson 1979), in which quotation is considered to be a special case of demonstrative reference (for a critical discussion of this view see, e.g., Harth 2011; Saka 1998; Washington 1992). More recently, the semantics and pragmatics of quotation have been receiving consistent attention. A central problem in theorizing quotation and the different meanings it can adopt is that the interpretation of a quotational construction heavily depends on the context in which the quotation is embedded. Consider the following examples.

The example in (4a) is an instance of a quotation that is used to inform the addressee about the linguistic shape of a certain concept’s conventionalized name, i.e., recursive function in this case. Quotational constructions of this kind typically involve naming predicates like call and have been analyzed as a special type of pure quotation (Cortés Rodríguez et al. 2022; Härtl 2018, 2020). In contrast, the quotation in (4b) adopts a distancing interpretation and, thus, supports an ironic reading of the expression in quotes, expert. A similar contrast can be observed between (5a) and (5b), where the quotation in so-called “thoracic outlet syndrome” in (5a) is interpreted in a name-informing sense, whereas the quotation in so-called “beach” again adopts a distancing, ironic interpretation. Irony-indicating uses of quotation like in (4b) and (5b) have sometimes been subsumed together with cases of metalinguistic quotation of the type in (4a) and (5a) as instances of scare quotation (e.g., Meibauer 2007; Predelli 2003).

Observe that the quoted nominals in the two examples in (5) above are used in a hybrid way. First, they denotationally function as the heads of the DPs a so-called “thoracic outlet syndrome” and the so-called “beach”, respectively. Second, the two nominals occur in a quotation, that is, they are mentioned, thus entailing that they are both used and mentioned at the same time. Recanati analyzes quotations where the expression inside the quotes contributes its denotation to the semantic content of the rest of the sentence as what he calls open quotations, e.g., A “fortnight” is a period of fourteen days. Closed quotations, in contrast, as in, e.g., “Fortnight” is an unfamiliar word, only contribute a metalinguistic referent to the compositional structure (Recanati 2001: 682–683).

2 Quotation and the semantic-pragmatic interface

3 The contributions

The special issue contains five contributions. The first three articles by De Brabanter, Finkbeiner and Schlechtweg and Härtl are written against the background or in support of pragmatic theories of quotation. In the paper Quotation does not need marks of quotation, De Brabanter argues for pragmatic theories of quotation and considers quotes to be merely optional markers with a disambiguating function, rather than an essential part of a quotation, as is assumed in semantic theories of quotation. He starts out with some general, but striking observations about the relation between quotation and quotes. It seems intuitively hard to imagine that quotes are more than just additional elements since marking quotation in a (relatively) systematic way was not established until thousands of years after writing had made its way into human interactions, a lack of systematicity had shaped the usage of quotes for a long time, and many of the world’s languages are only spoken, not written, which does of course not imply that in these languages quotation does not exist. The author supports a pragmatic approach called Depiction Theory where iconic meaning is at the center of interest, rather than the conventional pairing of form and meaning. In other words, what is placed between quotes is displayed. Quoting can still happen without quotes, and in these cases, the context, such as a metalinguistic predicate, can help the addressee detect the intended quotational reading. De Brabanter includes observations and findings from both written and spoken language in his discussion. For writing, one category of evidence he presents is from examples of pure and direct quotation which are unambiguously quotations but which lack any kind of marker, like quotes, to emphasize this status. Note that some examples without quotes are found even in sources where careful editing can be assumed. Hence, we find quotations without quotes, and these instances without quotes are semantically equivalent to the same examples with quotes. Looking at speech, the author first realizes that there is no systematic marking of quotation. Different researchers have found different acoustic strategies to implement quotation. This suggests that quotes in spoken language are optional, as otherwise, we would expect some kind of homogeneity in the expression of quotation. A second piece of evidence supporting this view comes from a recent study by Schlechtweg and Härtl (2020), which we mentioned in Section 2.2 and which found that language users systematically pronounce expressions within quotes differently from the same expressions without quotes; crucially, quotations without quotes are distinct from quotations with quotes, and quotations without quotes are more similar to cases that are clearly no quotations at all. Having reviewed these examples, De Brabanter continues tackling two alternative solutions to the problem: One might claim that quotations without quotes are in fact no quotations or that quotations without quotes are actually built on the basis of a quotative operator, under the surface so to speak. He shows that these alternative solutions are problematic and finally reflects upon how the Depiction Theory fits the picture he advocates.

In her article entitled Quotational nicknames in German at the interface between syntax, punctuation, and pragmatics Finkbeiner deals with a specific phenomenon where a nickname is positioned between the first and last name in a death notice, and this nickname is enclosed by either quotes or parentheses, e.g., Johanna „Hansi” Angermaier versus Ursula (Uschi) Vollmuth. Her objective is to contrast the usage of the two punctuation marks in the same context and genre in order to provide us with new insights into the function and meaning of quotes in particular and the syntax-punctuation-pragmatics interface more generally. Finkbeiner points out that the crucial difference between parentheses and quotes is that parentheses reflect the parenthetical structure of the construction and support the syntactic operation, while quotes signal a conversational implicature and the need to adjust the interpretation of the construction in the first place. The author begins by referring to the formal and functional commonalities and differences between quotes and parentheses. For instance, both represent twin signs with a pro- and an enclitic component that attach to a host, i.e., the material between them. The observation that one cannot simply interchange quotes and parentheses leads her to the discussion of the distribution of the two types of marks in the specific nickname construction in focus. The analyses are based on a corpus of death notices collected from several German newspapers. Finkbeiner finds, first of all, that the kinds of nicknames enclosed by quotes are linguistically more diverse than those detected in the examples with parentheses, that parentheses, but not quotes, offer the possibility of moving the nickname (without parentheses) to the front, with the first name following (being enclosed by the parentheses), and that expressions between quotes occur both before and after the full name (first plus last name) whereas constructions in parentheses always follow the full name. She then continues developing her broader theory and suggests that constructions with quotes and parentheses require a syntactic modification in comparison to the sequence of first and last name without any intervening nickname. More precisely, she interprets the data as cases of parentheticals, where a sentence or phrase is embedded into another sentence or phrase. This is the primary function of parentheses, with the nickname functioning like an inserted comment and with the parentheses being possibly replaced by a different type of parenthetical marker (e.g., en dash). In contrast, the primary function of quotes relates to meaning, not syntax. That is, an additional meaning aspect beyond the fact that the person used to be referred to by the nickname is pragmatically licensed and the addressee has to understand the conversational implicature triggered. An example of the additional meaning aspect is that a nickname enclosed by quotes is often not simply one used by everyone to address the person in focus; rather, it is a nickname used by people who were in a close and intimate emotional relationship with the person.

In the contribution Quotation marks and the processing of irony in English: Evidence from a reading time study, Schlechtweg and Härtl present psycholinguistic evidence in favor of pragmatic theories of quotation. They start by discussing the central concepts of the article, namely, quotation, quotes, and irony and foster the theoretical foundation of their research. The function and meaning of the process of quotation and a specific tool to emphasize this process, quotes, is examined first by paying particular attention to both semantic and pragmatic theories of quotation. The second important theoretical concept, irony, refers to situations in which an expression’s interpretation is non-literal and an alternate meaning is intended. This typically combines with a specific evaluation added by the respective speaker, thus expressing a negative attitude towards something. Quotes and irony are frequently found together in cases of scare quotation, which is the type of quotation that is in the focus of the empirical investigation. In scare quotation, the quotes can support the non-literal meaning and the deviation from the default denotation of an expression. In the next step, the state of the art of the psycholinguistic literature on irony and quotes is examined. In irony research, the (potential) interplay of the literal and non-literal meanings during language processing play a key role; in research on the processing of quotes, in turn, it has been analyzed whether quotes have acoustic correlates in production and whether they facilitate or inhibit the processing of quotation in perception. Schlechtweg and Härtl finally outline their own reading-time study in detail, in which they relied on data from 36 native speakers of English and were interested in finding out whether and how the presence of quotes influences the comprehension of ironic content in the written mode. To maximize comparability and reliability, they used the same sentences in the literal and ironic conditions and all sentences occurred both without (literal, ironic) and with quotes (literal, ironic). Only the preceding context was distinct across the conditions and triggered the intended interpretation. By measuring the reading times of different portions of the target sentences, the authors intended to disentangle potential effects and to determine under what conditions quotes have an effect on the processing of the sentences. It could be shown that quotes increase the processing burden at the moment of their appearance, and this applied to both literal and non-literal scenarios. Sentence finally, however, when individuals summarized the meaning of the entire sentence, quotes clearly facilitated the comprehension of irony. The authors use the fact that quotes do not facilitate processing at their direct occurrence but only later in time as one argument against semantic approaches to quotation.

Having presented three papers highlighting the pragmatic spirit of quotation, we now turn our attention to two contributions suggesting formal, semantic analyses of two specific cases of quotation. In Angry lions and scared neighbors: Complex demonstrations in sign language role shift at the sign-gesture interface, Steinbach observes that pragmatic and semantic theories of quotation have been mostly based on examinations of spoken and written language. He uses this as a point of departure for his studies of quotation as expressed in sign language and argues that these types of analyses are essential for a comprehensive modeling of the phenomenon. The author takes a recent approach to American Sign Language outlined in Davidson (2015), extends and adjusts the proposal, applies the result to German Sign Language, and shows how one can more fully account for a complex process by considering gestures, linguistics, and different body parts, all of which are at play in role shift, for example, to represent and report on various people simultaneously. He relies on cognitive linguistic investigations and finally offers a formal analysis of complex demonstrations in role shift. In sign languages, three characteristics are of particular importance in the context of role shifts. First, multiple articulators, the dominant hand, the non-dominant hand, the arms, the head, the face, and the upper part of the body, operate simultaneously to convey multiple meanings simultaneously. Second, the modality used in sign languages is the same as the one used in manual and non-manual gestures, which, in turn, explains the manifold interactions of sign language and gestures. Third, sign languages have a three-dimensional signing space at their disposal, which is used to articulate many linguistic aspects, such as agreement or coordination. He considers two types of role shift, the attitude/quotational one and the action role shift/constructed action. The former connects features of both direct and indirect speech and reports linguistic actions (e.g., a speech act); the latter often comprises a gestural demonstration of someone’s non-linguistic action (e.g., the action of walking). The author reminds us that the two types of role shift often go hand in hand by implementing comparable markers and by integrating linguistic and non-linguistic, gestural elements. To achieve his goals and establish his hybrid theory unifying attitude and action role shift, Steinbach uses two fables that a native signer translates into German Sign Language. These fables are appropriate to illustrate the complexity of quotation in sign language and enable the author to deepen our insights and foster available theoretical models with evidence from outside spoken and written language.

In his article Scare quotes as deontic modals Wiślicki aims at suggesting a formal semantic account of one particular type of quotation, namely, scare quotation. He observes right at the beginning of his work two open issues in the field, namely, first, that we still lack a well-defined and precisely formulated formal operation explaining the semantic features of scare quotation, although many researchers agree that the negative evaluation associated with scare quotation represents content that is not at issue, and, second, we still need a well-motivated explanation for why certain cases of scare quotation are ill-formed. The author demonstrates the complexity of the type of quotation under analysis and reflects upon it throughout the paper, eventually proposing an account in which scare quotation is regarded as a case of deontic modality. Quotation and more specifically scare quotation is an instance of modality and these modal characteristics lead the author to his model. Wiślicki further underlines that the negative evaluation, a key aspect of scare quotation, and different restrictions of this type of quotation overlap with violations of norms that become evident when producing the quotational construction. While developing his semantic theory, he refines and re-explains some of the properties of scare quotation. The negation of the literal content inherent to the quoted expression has been discussed in the context of ironic content but Wiślicki shows that this is not necessarily the whole story, rather, rejecting specific parts of the content, at-issue content, but also not-at-issue content (e.g., presuppositions), and opposing to the appropriate use of the construction/distancing oneself from the use of the construction seems to be at play and characterizes the concept of scare quotation more adequately. Besides semantic effects, the author also detects effects in other domains, for instance, cases in which the rejection is syntactically or pragmatically motivated. The analysis illustrates the flexibility of scare quotation but also incorporates its restrictions in a subsequent step. These limitations include the requirement that an expression be meaningful, the blocking of co-referentiality, and the tendency of ironic content to target not-at-issue content. Wiślicky integrates the significant interaction between a quoted construction, the context, and the common ground shared by the interlocutors into his discussion.