3.1 Ambiguity in Linguistics

A first kind of definition which is frequently found in linguistic approaches states that ambiguity is the

In other words, ambiguity is the capacity of a sign to convey distinct meanings:

According to Kennedy, these distinct meanings involve variability in the truth conditions of the utterance. Ambiguity is thus seen as a kind of interpretative uncertainty which

More specifically, ambiguity can be found in different types of linguistic entities (cf. the first definition above), and we frequently find an opposition of two basic types here, namely lexical and syntactic (or structural) ambiguity.¹⁶² While lexical ambiguity is confined to the meanings of single words (cf. example (13)), syntactic ambiguity arises from the possibility of different syntactic interpretations of a sentence (cf. example (14)). This means that ambiguity can be detected both within syntax / grammar and within the lexicon of a given language.¹⁶³

(13) bank – ‘financial institution’ / ‘edge of a river’

(14)

A well-known type of syntactic ambiguities are scope ambiguities, as illustrated by example (15), which can be interpreted with a wide scope reading of the universal quantifier (a.) or a wide scope reading of the existential quantifier (b.).

(15) Everyone loves someone.

1. ∀x [person (x) → ∃y [person (y) ∧ love (x,y)]] /
2. ∃y [person (y) ∧ ∀x [person (x) → love (x,y)]]

Another specific case of syntactic ambiguities are garden path sentences which mislead the hearer to a false interpretation; s/he then has to backtrack and look for an alternative syntactic interpretation. For example, in (16) the initial interpretation will probably be ‘the old man = the man who is old…’, but the hearer is then forced to reinterpret the sentence as ‘the old (persons) man the boat = the boat is manned by the old’.

(16) The old man the boat.

Normally, ambiguity is only a preliminary phenomenon in garden path sentences, and similar cases of ambiguity are treated as local ambiguities (or temporary ambiguities) and opposed to global ambiguities which remain after the entire sentence has been uttered (cf. examples (14) and (15)).

Syntactic ambiguities can equally arise by ellipsis.¹⁶⁴ In (17), one possible interpretation of the sentence would be that it is about ‘women more beautiful than Angelina Jolie’, another one, with a different scope reading, would be in the sense of ‘more beautiful women than just Angelina Jolie’. But the sentence could also be interpreted as ellipticalin (‘more beautiful women than Angelina Jolie does’).

(17) I know more beautiful women than Angelina Jolie.

Some types of ambiguity, however, seem to go beyond the clear-cut distinction between lexical and syntactic ambiguity, as the lexical and syntactic level may interact. For example, in (18) an ambiguity arises with respect to the meaning of both can and fish, the latter form being interpretable as a noun or as a verb.

(18) They can fish. – ‘they may or are able to fish’ / ‘they put fish in cans’

Interestingly, an interpretation of can in the sense of ‘to be able to’ already determines the interpretation of the second, potentially ambiguous form as well. This means that the resolution of lexical ambiguities has an impact on the syntactic interpretation of the sentence, and syntactic aspects contribute to a disambiguation of lexical ambiguities.

Let us now turn to lexical ambiguities. A parameter of analysis which is central to traditional approaches is the semantic distance between the two meanings of an ambiguous term. According to this parameter, homonymy and polysemy can be distinguished (cf. Blank 1997, 417, 2003, Bauer et al. 2010, 48, Bode 1997). In the case of homonymy, there are no immediate semantic relations between the two meanings (cf. example (19)), and we deal with two signs being independent from each other and homonymous “by chance”.¹⁶⁵ In polysemy, by contrast, the two meanings are clearly distinct, but at the same time semantically related (for example, in (20) there is a metaphorical relation between the two meanings). Moreover, these two types of ambiguity can be opposed to cases of vagueness, where the distance between the meanings is minimal, so that it is difficult to treat them as clearly separated meanings (cf. example (21), where the word tall is vague, insofar as its interpretation depends on various contextual factors such as the age of Jenny, the subject of the discussion and so on. Is Jenny a 3-year-old or an adult person? Is she a basketball player or a jockey? etc.). Vagueness is therefore distinguished from ambiguity and treated as a distinct phenomenon (Pinkal 1991, see also Bode 1997, 68).

(19) E. bank ‘financial institution’ / ‘edge of a river’

(20) E. mouse ‘mouse’ / ‘computer mouse’

(21) Jenny is tall.

The examples that have been discussed so far share some basic features: they illustrate ambiguities which exist in the language system, and they are potentially disturbing in communication, which means that a process of disambiguation is required. Therefore, a central aim of previous studies is to determine how disambiguation takes place (cf. Fass 1988). If contextual clues do not permit the hearer to interpret the ambiguous term or structure in accordance with the meaning intended by the speaker, communication may be obstructed. At the same time, the various categories which have been developed in linguistic approaches show that ambiguity is a frequent phenomenon which turns up in different ways in everyday communication.

3.2 Ambiguity in Literary Studies

In literary studies, the fundamental importance of ambiguity is put forward in an even more explicit manner, as ambiguity is regarded as a basic characteristic of literary texts. A general view is that, contrary to the aim of disambiguation in a linguistic perspective, literature seeks to be ambiguous (cf. Wachter 2006, 177–178); hence, reflection on literary communication and texts has been linked to the concept of ambiguity (cf., e.g., Eco 1989, Bode 1997, Kronauer 2002, Furniss & Bath 2007, Berndt & Kammer 2009, Bauer et al. 2010), and ever since the studies by Empson (1930) and Jakobson (1960), ambiguity has been regarded as one of the “roots of poetry” (Empson 1930, 113, cf. Bauer et al. 2010, 28). In the context of literature and literary studies, ambiguity is therefore regarded as an indicator of aesthetic creativity (see Bauer et al. 2010, 28) and of “formal innovation in art” (Robey 1989, xi).

Jakobson in particular made a distinction between ambiguity in literary/ poetic texts as an extraordinary case of communication in opposition to ambiguity in ‘ordinary communication’.¹⁶⁶ The particular language use in literature is what he calls poetic function (Jakobson 1960, 356–357).¹⁶⁷ He regards self-referentiality, i.e. the equivalence between different linguistic (and phonetic) signs within the utterance, as the characteristic feature of poetic diction, both with regard to “phonological sequence” and “semantic units” (Jakobson 1960, 370): “The poetic function projects the principle of equivalence from the axis of selection into the axis of combination. Equivalence is promoted to the constitutive device of the sequence” (Jakobson 1960, 358, emphasis original; see also Bauer et al. 2010, 28). Through this, the reference becomes ambiguous: “Ambiguity is an intrinsic, inalienable character of any self-focused message” (Jakobson 1960, 370–371). These reflections lead to his conclusion that, in literary language and communication, marked by their poetic function, the equivalence of signs is more important than reference, i.e., that form is more important that content.¹⁶⁸ In that a message refers to itself, it is embedded into communicative contexts to a lesser extent and therefore more open to interpretation and ambiguity: “The supremacy of poetic function over referential function does not obliterate the reference but makes it ambiguous” (Jakobson 1960, 371).

Christoph Bode in his seminal study on Ästhetik der Ambiguität (1988) follows Eco in making a connection between ambiguity and the modernity of a text.¹⁶⁹ This view is insofar problematic as it disregards the multiple use of ambiguity in earlier periods, i.e. before modernism; during the Early Modern period, for instance, ambiguity was widely used in poetry (and drama), e.g. in the works of William Shakespeare, John Donne, George Herbert etc. What he notes, much in the wake of Jakobson, is a particular quality of poetic language that deviates from daily speech (Bode 1988, 25), the essential ambiguity of poetic texts (Bode 1988, 71).

While Bode seems to include polyvalence and ambivalence as well as indeterminacy under the heading of ambiguity, there are other literary critics that have chosen a more narrow scope and regard ambiguity strictly in terms of duality. In her book on Henry James, Shlomith Rimmon (1977) follows such an approach and focuses mainly on the simultaneity of two mutually exclusive meanings.

In their contribution on “Dimensions of Ambiguity”, Bauer et al. (2010) take into consideration the production of literary ambiguity and distinguish between various kinds of ambiguity, stressing, however, that these are usually linked with one another. They notice that ambiguity may arise when several communicative levels come together. A further source of ambiguity is the evocation of contexts, for instance in Jane Austen’s Mansfield Park, where the question arises whether Sir Bertram is a slave owner or not, i.e. a good character or a bad, as the novel only mentions his owning plantations in Antigua.¹⁷⁰ A third source is linguistic self-referentiality (see Bauer et al. 2010, 31–35). In this sense, according to Bauer et al. ambiguity can affect parts of texts or whole works, very much in the sense of manifold interpretation of the Scriptures¹⁷¹; ambiguity can be the object of literary mimesis; and lexical, semantic and syntactic ambiguity can become functional in literary texts (Bauer et al. 2010, 35–38).

What is perhaps most relevant in the present context – besides the interdisciplinary approach also chosen by Bauer et al. – is the recognition of various communicative levels that come into play in a literary text and cause ambiguity, as in example (22) from Romeo and Juliet:

(22)

Romeo’s statement, as Bauer et al. (2010) note, is ambiguous: not only does his “O dear account” refer to the nurse’s statement about riches and wealth, “the chinks”, but also does he refer to the high price he will have to pay for his love; what is more, he also says, at the same time, that the news is dear to, i.e. cherished by him as he loves Juliet already after their first meeting (see Bauer et al. 2010, 29). What follows for Bauer et al. is that, although literary communication may be regarded as a particular case in point, it still has to obey the rules of ordinary communication that is imitated in drama (in accordance with the notion of mimesis as expressed in Aristotle’s Poetics). This purports that, despite the fact that literary / poetic communication may be regarded as a special case of communication and is essentially aesthetic (cf. Jakobson 1960, 371), it is not necessarily less comprehensible than other cases of communication. The aesthetic nature of a text does not imply that an author is exempt from the need to express himself understandably –but as aesthetic communication is full of ambiguities it may here even become more difficult to distinguish between intended or strategic ambiguity and accidental ambiguity (cf. Bauer et al. 2010, 31).¹⁷³ The message uttered by characters in a text may take on meanings for the reader that escape the characters involved in communication within the text, as in (23):

In Shakespeare’s As You Like It, Rosalind disguises as a man when she escapes with her cousin Celia to the forest of Arden. Her lover, Orlando, has fled to the same forest, and there they are now playing a strange game: he does not know that the youth Ganymede whom he meets is in fact Rosalind; Rosalind, on the other hand, wants him to behave as if she were Rosalind although Orlando thinks she is Ganymede (see also Brown 1990). In the course of events, Orlando fights a lion and is hurt. This piece of news is now delivered to Rosalind – still in disguise –by Oliver (Orlando’s brother), who shows her a bloodstained handkerchief. She faints. Oliver’s statement, “You lack a man’s heart”, alludes to Ganymede’s youth and mocks him because of his weakness. Rosalind’s answer, however, is to be understood as a reaction that takes the statement literally – all this is lost on most of the (other) characters, who do not know about her role-play. For the audience, this exchange is great fun as they share Rosalind’s knowledge. Through dramatic irony, the statement becomes polyvalent.

Reboul (2001), following Banfield (1982) (see also Bauer et al. 2010, 32) addresses these several levels of communication and the ambiguity resulting from them, with a particular focus on narrative texts.¹⁷⁴ She explains that, apart from the interaction of communicative partners within a narrative (or drama, for that matter), there is another level going beyond that, namely that of narrator and reader (or, in the case of drama, of author and audience, as in (23)). According to her, this results, again, in ambiguity (cf. Reboul 2001, 254): a sentence may be unequivocal and perfectly unambiguous and its linguistic meaning may be unchanged, but still the utterance stated in different contexts may have different meanings (see Reboul 2001, 257). This is particularly the case when “it is not obvious who is speaking” (Reboul 2001, 257); ambiguities resulting from this she calls “pragmatic ambiguity” (Reboul 2001, 254). It can be found, for instance, in narrative texts when it remains unclear whether a sentence is uttered by the narrator or is an example of free indirect discourse, i.e. its meaning depends on the level of discourse (cf. Bauer et al. 2010, 32). Reboul chooses an example from Dickens’s Great Expectations in order to illustrate this point:

(24)

She interprets this passage in the following two different ways (Reboul 2001, 262):

(25)

(26)

The first reading is based on interpreting the passage as if told from the narrator’s point of view – and as Great Expectations is an autodiegetic novel, meaning that the narrator talks about his own life, this would be the thoughts of Pip in the past, while “the rest of Pip’s story shows that Miss Havisham was not in anyway his benefactor and that he was mistaken in thinking such a thing” (Reboul 2001, 262), which means that (25) is actually the rendering of (24). The second reading follows the notion that this is the retrospective narrator’s true statement. The ambiguity based on the dialectic between experiencing and narrating I is a standard problem in autodiegetic narrative – as is the dichotomy between narrator’s voice and free indirect discourse (see also Bauer et al. 2010, 32). In Great Expectations , this means that, in the course of reading the novel, an ambiguity – is Miss Havisham Pip’s benefactress or is she not – is done away with, and the reader becomes aware of this. An existing ambiguity is therefore resolved (see 4.2). Thus, the processing of messages on part of the reader comes into focus.

3.3 Ambiguity in Other Disciplines

Our overview of traditional approaches to ambiguity in linguistics and in literary studies has revealed important differences in the ways in which ambiguity is perceived and defined. While in many traditional linguistic analyses, the assumption prevails that the central function of language is the efficient transmission of information – which entails the idea that ambiguity potentially hinders successful transmission – literary approaches stress the creative aspects of ambiguity, as ambiguity permits a speaker (or author) to suggest different interpretations and invites the hearer (or reader) to reflect on possible interpretations of a given message (or literary text). We would like to conclude our overview by briefly looking at traditional approaches to ambiguity in other disciplines in order to find out if the approaches mentioned above cover all aspects of ambiguity or if further aspects of ambiguity are taken into account by other disciplines.

For many disciplines, either the traditional linguistic conception of ambiguity or the approach developed in literary studies seems applicable. For example, in theology, an important field of studies where ambiguity becomes central is the fourfold (or manifold) interpretation of the Scriptures. The possibility of multiple interpretations of a given text reminds us to a certain degree of literary uses of ambiguity.¹⁷⁵

In the study of law, it can be assumed that the central functions of communication are the successful “transmission” of information in legal cases and the binding force of legal documents: on the one hand, legal cases are usually formulated in a way which aims at avoiding ambiguity (no further interpretations beyond the intended description of the facts of a case should be permitted; cf. Thaler 1982), but on the other hand – for example in the context of international laws (cf. Grewendorf & Rathert 2007, Rathert 2006, esp. 82–86, Kötz 2003, Pehar 2001, Thümmel 1998) – certain ambiguities can also be intended in order to permit that the text of a certain law can be applied to the specific situations in different countries.¹⁷⁶

In rhetoric, the study of ambiguity goes back to antiquity.¹⁷⁷ To Cicero and Quintilian the issue of ambiguity arises in the context of quarrels, controversies and legal disputes (see Bauer et al. 2010, 11). In these contexts, ambiguity is mostly regarded as a flaw, a vitium which must be avoided (Quintilian 1966, 8.2.16). When ambiguity is used to obscure things (as, for instance, in the oracle), Quintilian condemns it and demands clarity (perspicuitas) of expression (Quintilian 1966, 8.2.16–22). Strategic uses of ambiguity, however, may be justified, e.g. when the orator wants to avoid obscenity and rather uses a double entendre, or when an unequivocal utterance is unapt, e.g. when the orator criticises a tyrant (cf. Quintilian 1966, 6.3.47–49, see also Bauer et al. 2010, 24, Bernecker & Steinfeld 1992, 437).¹⁷⁸ The study of rhetoric therefore distinguishes between ambiguity in ordinary communicative settings, in which there is a tendency towards disambiguation in order to achieve a decision or understanding, and ambiguity in special cases of communication such as literary texts where it becomes aesthetically productive (cf. Bauer et al. 2010, 9, see also Wagner-Egelhaaf 2009). Moreover, the rhetorical approaches hint at a complex interplay of the speaker’s and hearer’s actions in communication. In fact, rhetorical arguments mainly address the (anticipated) reaction of the hearer and then formulate instructions that the speaker should follow accordingly. This means that speaker-hearer-interaction in communication is not only a “linear” process, but that at each step, in planning and interpreting the utterances, both participants also reflect on the attitudes, intentions etc. of the communication partner.

This leads us to psychological studies, where a somewhat different conception of ambiguity has been developed. An important model for psychological approaches to ambiguity is the Donald paradigm which is used to study judgmental priming effects (cf. Higgins et al. 1977, Srull & Wyer 1979, Ziegler 2010, 134). In approaches following this paradigm, participants are faced with an ambiguous description of a character named, e.g., Donald (see example (27)). In a second step, participants are then asked to judge this character. It has been shown that priming effects can be observed in their judgments, that is, if participants first perform a sentence construction task which activates concepts associated with either hostility or kindness (e.g. break, leg, arm vs. hug, kiss, cf. Srull & Wyer 1979, 1663), they will later be inclined to judge a character as being more hostile or kind based on a description of his/her behaviour which permits both interpretations (cf. Srull & Wyer 1979).

(27)

Other studies have focused on the persuasive effectiveness of arguments according to parameters such as the source of the utterance, recipients’ attitudes, etc. (cf. Ziegler & Diehl 2003, Ziegler 2010, 137–140). Here, the term ambiguity is used to characterize certain arguments which have, if presented in isolation, an intermediate degree of persuasiveness¹⁷⁹ and thus permit different interpretations in the sense of an evaluation as a rather strong or a rather weak argument by the recipients. In example (28), the test subjects were faced with an argument that was presumably uttered by one of the two candidates for the office of Federal Chancellor in the election of 2002 (Gerhard Schröder and Edmund Stoiber), and it has been shown that the test subjects’ interpretation of this argument as a rather strong or a rather weak one depended on the political preference for Schröder or Stoiber.

(28)

This brief overview shows that psychological surveys focus on a specific kind of ambiguity. In the experiments, ambiguity is produced by a certain utterance, but we cannot identify specific parts of the utterance which are ambiguous, as it is the entire utterance which can be interpreted differently. From a linguistic perspective, we could thus characterize these examples as cases of pragmatic ambiguity, the context (the primes) being a predictor for the positive vs. negative judgment. At the same time, in the Donald paradigm, it is ultimately the character him/herself who is judged, so that still another aspect becomes relevant here, namely the interpretation of extralinguistic reality. We will see in section 4.1 that this aspect also plays a role in a genuinely linguistic perspective, when we focus on cases of onomasiological variation / reanalysis, where one and the same part of reality can be interpreted differently. Summing up, in psychology, ambiguity is approached in terms of everyday situations and attitudes. At the same time, it is related to language phenomena as it frequently implies the production of an utterance and its perception by a hearer. The interaction between speaker and hearer therefore seems to be a basic prerequisite when it comes to the study of ambiguity. However, this aspect has, surprisingly, not really been studied extensively so far.

3.4 Towards a Redefinition of Ambiguity in Speaker-Hearer-Interaction

The various examples of ambiguities analyzed in linguistics, literary studies and other disciplines show that the concept of ambiguity itself can be seen as ambiguous. Approaches rooted in the tradition of logic and semantics focus on the successful transmission of information and the correct identification of the referent, which means that the focus is mainly on processes of disambiguation. These approaches define the term in a relatively narrow sense: depending on the interpretation of the ambiguous term or syntactic structure, one and the same utterance can be assessed as either true or false. The two interpretations of the ambiguous term or structure are clearly distinct, as they represent different lexicalized senses of a word or different syntactic structures which can be assigned to a given sentence. There is a semantic focus on ambiguity, and ambiguity is primarily analyzed as a potential of abstract linguistic signs and structures which are part of an overall linguistic system.

At the same time, ambiguity is frequently studied from the perspective of the hearer’s interpretation in these approaches. It is assumed that, for the hearer, ambiguous utterances are potentially difficult to process, as the identification of the correct interpretation may cause increased processing efforts. This perspective on ambiguity ties in with the tradition of rhetoric which regards ambiguity as a vitium (however, we have seen that rhetorical approaches also take into account strategic uses of ambiguity in specific circumstances, considering a complex interplay of speaker-hearer-interaction and mutual anticipations), and, ultimately, in many traditional analyses, the focus is thus on disambiguation.

Psychological approaches, in contrast, adopt a much wider perspective: ambiguity is no longer related to single parts of utterances, but manifests itself as a characteristic of entire utterances or situations. Moreover, in many cases, ambiguity does not designate an alternative between two (or any identifiable and restricted number of) clearly distinct interpretations, but the different interpretations which can be assigned to the ambiguous utterances are located along a continuum (ranging, e.g., from a markedly negative or a slightly negative to a slightly positive or markedly positive judgment etc.). At the same time, the ambiguities studied in psychological approaches are mostly non-strategic (at least, on a first level of communication – on the level of scientific analysis, of course, the psychologist’s use of ambiguous descriptions or arguments is completely deliberate).

In literary studies, ambiguity is also frequently analyzed on the textual level, that is, beyond the level of single words or syntactic structures, and the essentially ambiguous quality of poetic texts – including poetry, narrative texts and drama – is being noted and stressed. Ambiguous statements are almost always intended by the writer of a literary work. In the majority of cases, the ambiguity is not addressed to the speakers and hearers within the literary work but rather directed at a reader or an audience. In these cases, communicative processes are thus represented in literary works, and this is why several communicative levels come into play here.

The examples of ambiguity analyzed in literary studies are of great interest for general approaches to ambiguity, as literary texts exhibit intentional uses of ambiguity which are very complex and sophisticated and which involve various levels of communication. Moreover, beyond the actualization of ambiguities in the language system, literary texts also make creative use of the repertoire of signs given in the language and in this way introduce new ambiguities. This creativity becomes apparent in different ways. On the one hand, the actualization of the signs of the linguistic system in their conventional meaning involves a creative element, insofar as the speaker selects and combines these signs in a novel way, thus potentially creating new associations and, in some cases, inviting the hearer to explore the possible semantic interpretations arising from the conventional meaning. As an example, we could give the title of Ngaio Marsh’s detective novel Scales of Justice (see below example (55)). By combining the words scale and justice, the speaker suggests an interpretation in the sense of ‘the balance of Justice’. However, the text of the novel then suggests a further reading which is based on another conventional meaning of scales, namely ‘the scales of a fish’. On the other hand, we can also observe cases where the speaker arranges his/ her utterance in such a way that s/he suggests a novel interpretation which goes beyond the conventional meaning(s) of the words that are used, and invites (or forces) the reader to semantically reinterpret the word in order to make the utterance become coherent.

However, creativity does not only play a role in literary texts, but it is, in both senses described, also an important feature of linguistic activity in general, that is, creativity can be observed in everyday language which is characterized by a sort of “everyday rhetoric” (Stempel 1983). Especially with respect to the creation of new ambiguities in discourse, it therefore also seems promising to reconsider previous approaches to language change (cf. section 2). While it is now well recognized that semantic change typically involves a stage of polysemy / ambiguity in the language system (where the old and the new meaning coexist), it is noticeable that the introduction of new meanings at the stage of innovation – which can involve a situation of ambiguity in discourse – has only rarely been analyzed under the label of ambiguity in linguistic studies, and traditional definitions of ambiguity are restricted to cases where the ambiguity already exists in the language system.¹⁸¹

Bringing together the two perspectives just mentioned – literary studies and language change – in an interdisciplinary approach, the main question concerning the study of ambiguity can thus be reformulated in the following way: instead of asking why ambiguity should be avoided (or why in some cases it need not be avoided, cf. Wasow, this volume), we propose to analyze the realization of various types of ambiguity in communication and to enquire why ambiguity should be used and introduced by speakers and hearers.

In order to address this question, we adopt a pragmatic approach to the study of ambiguity. Focusing on ambiguities on the level of discourse, we propose a concept of ambiguity in speaker-hearer-interaction (cf. Winter-Froemel & Zirker 2010), and we will discuss key parameters of analysis which can be applied to phenomena of ambiguity across the disciplines. Based on the assumption that ambiguity can be seen as a fundamental characteristic of human utterances, we will emphasize the roles of the communication partners – speakers and hearers –in analyses of ambiguity, and consider not only the resolution of ambiguities (or their disambiguation), but also focus on the creation of ambiguities in communication. What is more, we will take into account very different types of ambiguity and their potential communicative advantages as well.

At the same time, we have chosen an interdisciplinary approach. Instead of staying within (artificial) disciplinary boundaries – “the linguist uses an eitheror method, the critic [in literary studies] a both-and procedure” (Hagenbüchle 1984, 216) when it comes to analyzing ambiguity –, we propose to stress the factor that unites both disciplines, namely language. Contrary to the opinion that there is a “fundamental difference in their attitudes to language” (Hagenbüchle 1984, 215), we will argue that both linguistics and literary studies provide tools that enhance each other and, in this combination, lead to new findings that go beyond methods of analysis and interpretation that are restricted to the respective fields. Ambiguity is a field of studies that interests both linguistics and literary studies (and many other disciplines, as this volume shows) and that is especially apt to try and unite the perspectives of the fields (see also Bauer et al. 2009, 253–254) because thus a more comprehensive set of specific parameters of analysis can be determined.

4.1 The Linguistic Material: Form and Meaning / Interpretation(s)

A first set of parameters can be proposed with respect to the ambiguous message itself. As shown in section 3.1, previous research in linguistics has mainly focused on this aspect of ambiguity, distinguishing between lexical and syntactic ambiguity. However, we have already seen above (cf. sections 3.1, 3.2, 3.3 and 3.4) that ambiguities can also turn up on other levels of linguistic analysis, such as the morphological and the textual level. Moreover, we can also observe ambiguities located within the extra-linguistic reality, i.e. in the objects designated with regard to their conceptualization, and these ambiguities may be reflected in language in cases of semantic change.

In order to sort out the various possibilities that can be realized here, it is helpful to refer to Morris’s subdivision of semiotics into syntax, semantics and pragmatics (Morris 1946, 217–220, 1971, 21–24 and 301–303). According to Morris, the domain of syntax is concerned with the combination of signs, whereas the domain of semantics analyzes signs and their signification, and the domain of pragmatics studies the origin, uses and effects of these signs in communication, i.e. their relation to the speakers. It seems problematic to use Morris’s labels in order to classify subtypes of ambiguity, as all manifestations of ambiguity have a semantic component, so that a term like “semantic ambiguity” seems to be illfounded. Nevertheless, the three domains can help us to identify three clearly distinct kinds of ambiguity.

First, we can analyze cases of ‘atomic ambiguity’ (cf. the domain of semantics according to Morris) that are concerned with individual signs which can be assigned different interpretations (cf. example (29)). The ambiguous entities which are analyzed here can be lexical elements (cf. the traditional concept of lexical ambiguity), but also morphemes (cf. example (30) which shows that the Italian morpheme {-e} is used to express the masculine or feminine singular or the feminine plural).¹⁸² Cases of ‘compositional ambiguity’, in contrast, are characterized by the fact that the identification of the various (individual) elements of the utterance is at stake here (cf. Morris’s concept of syntax). This means that we can include classical examples of syntactic ambiguity, as given in (31), where the meanings of single words remain unchanged in both possible syntactic interpretations. Similar processes can also occur on lower levels of analysis, as given in cases of agglutinations or deglutinations (cf. the agglutination of the definite article in the name of the French town Lille, see example (32)) and in playful resegmentations of a certain word (see example (33)). Still another kind of ambiguity can be observed in cases where the entire utterance has to be considered as being ambiguous and where various possibilities of interpretation arise depending on the context. This latter kind of ambiguity can be linked to Morris’s concept of pragmatics. As we have seen in example (23) above, this kind of ambiguity appears in literary communication, but it can also arise in everyday communication (see example (34) and the domains of indirect speech acts and pragmatic implicatures in general). For instance, utterances like in (34) can be interpreted either just literally, that is, as assertives, or simultaneously as indirect speech acts (in this case, a request or, according to Searle, a directive; cf. Searle 1976).

(29) The men decided to wait by the bank. (cf. Simpson 2004, 367)

(30) I. nazion-e_FEM.SG, fior-e_MASC.SG vs. donn-e_FEM.PL, ragazz-e_FEM.PL

(31)

(32) F. Lille < l’isle ‘the island’

(33)

(34)

From a semiotic perspective, these three kinds of ambiguity share the characteristic of adopting a semasiological perspective: for given elements of expression, different interpretations can arise. In addition to these cases, a fourth possibility is given by cases of onomasiological ambiguity: here the perspective is reversed, and the starting point are the objects of the extra-linguistic reality, which may be perceived and conceptualized differently and consequently expressed by different linguistic means (see example (35) which illustrates a shift from a collective interpretation to a holistic one, and example (36), which illustrates a change in the opposite direction, cf. section 4.4). This means that the ambiguity is not primarily located in linguistic structures here, but in the objects designated themselves. This kind of ambiguity, however, is of interest to our concerns to the extent to which it potentially affects language, giving rise to semantic innovations and semantic change.

(35) Lat. ligna_NEUTR.PL PIECES OF WOOD → ligna_FEM.SG WOOD

(36) G. Trauben BUNCHES OF GRAPES → GRAPES

Based these basic options, we can therefore formulate a first parameter of analysis:

Ambiguities can thus arise on different levels of analysis. We have already seen in section 3.1 that, in some cases, the various categories of ambiguities mentioned in P1 may overlap. For instance, a message may show both an atomic and a compositional ambiguity (cf. example (37)). Moreover, within atomic ambiguity, different levels of analysis can be concerned, as given in (38), where the morphological ambiguity of the morpheme {-e} in Italian appears together with a lexical ambiguity (between the masculine noun signore ‘Mister’ and the feminine noun signora ‘Mrs.’). Even here, however, in many cases of use the interpretation will be disambiguated by the context (e.g. “signore e signori” ‘ladies and gentlemen’ disambiguates towards the feminine plural reading, “Egregio signore” ‘Dear Sir’ disambiguates towards the masculine singular reading, etc.), so that the hearer will easily identify the correct (intended) meaning, possibly without noticing the potential ambiguity of “signore” at all (cf. the parameters P4, P6 and P8 defined below).

(37) They can fish. – ‘they may or are able to fish’/‘they put fish in cans’

(38) I. signore_MASC.SG (of signore) – I. signore_FEM.PL (of signora)

Another example of ambiguity which involves an atomic and a compositional component is given in example (39), taken from Lewis Carroll’s Alice’s Adventures in Wonderland. The NP is as a rule interpreted with turtle soup being the head and mock being the modifier. However, Lewis Carroll plays here with the potential ambiguity of this structure: according to the syntactic rules of the English language, the interpretation with the head soup and the modifier mock turtle is equally possible, and, following this syntactic reanalysis and rebracketing, the name Mock Turtle is given to a creature that is made up of a calf (its head and tail) and a turtle (see the illustration by Tenniel, in Carroll 1998, 89).

(39) mock turtle soup [mock] [turtle soup] / [mock turtle] [soup]

Carroll not only plays with the ambiguity of the compound mock turtle soup, but goes further in that he takes the notion literally and invents a creature based on the ambiguity contained in the expression (cf. Zirker 2010, 234).

For cases of atomic ambiguity, another parameter can be defined depending on the existence or non-existence of a semantic or cognitive relation between the two meanings involved. These two options can be associated with the traditional concepts of polysemy and homonymy¹⁸³:

An example of an ambiguity involving homonymic forms is given in (40). Example (41), in turn, illustrates a case of ambiguity where the two meanings ‘funny-ha-ha’ and ‘funny-peculiar’ are interpreted as semantically related.

(40) The men decided to wait by the bank.

(41) He is funny.

Moreover, for cases of polysemy / semantically related meanings¹⁸⁴, we can further investigate the possible semantic relations between the two meanings involved (see parameter P3). Previous research has shown that there are seven types of relations which can occur (cf. Blank 1997, 2003, Koch 2000, 2001). The set of relations is universal, that is, it is applicable to all languages, thereby permitting an exhaustive analysis of all possible types of polysemies.

According to this set of relations, the two senses of funny in (41) could be interpreted as being related by contiguity, and the two meanings of mouse, ‘animal mouse’ and ‘computer mouse’ (cf. sections 2 and 3.1), show a relation of metaphorical similarity.

4.2 Ambiguity in Discourse: Perception, Strategic Use, Disambiguation

The parameters above have been discussed for the linguistic entities concerned, independently from their use in communication, and it thus seems possible to apply them to ambiguity in the language system as well as to ambiguity in discourse. The following set of parameters, in contrast, explicitly focuses on the level of discourse and on the actualization of ambiguity in individual situations of communication.

A first parameter of analysis here concerns the question whether the ambiguity is perceived by the speaker and hearer of the utterance, or whether they simply miss it when producing and interpreting the message.

The latter option is illustrated by the following example, which has already been cited above (cf. section 2):

(42)

It is clear that most readers would not even suspect an ambiguity on the discourse level here and choose the meaning of a ‘financial institution’, which is by far more plausible from the context. Theoretically, however, the word bank could be seen as ambiguous, as it has two different meanings on the level of the language system and neither meaning is explicitly ruled out in the utterance. From a psycholinguistic perspective, it is also very likely that the meaning ‘financial institution’ is activated by the previous mentioning of the word money. In this sense, the hearer will most probably not perceive an ambiguity here, as the meaning of bank is immediately disambiguated.¹⁸⁶ Another example of a non-perceived ambiguity is given in cases where L2 language learners fail to understand a wordplay due to their incapacity to identify the second meaning evoked in a certain context:

(43)

In the following example (cf. section 2), we can equally assume that the hearer does not perceive an ambiguity. This time, however, the two meanings are relatively close to each other, so that the difference of the possible interpretations is negligible in communication:

(44)

Other examples which have been analyzed above in turn illustrate cases where the ambiguity is clearly perceived by the hearer. This would be the case for (43) above, given that the hearer understands the wordplay contained in this sentence. Another example in this context is the garden path sentence presented in section 3.1 (cf. example (45)): here, the hearer is faced with a problem in language processing and, in order to solve this problem and to obtain a grammatical understanding of the sentence, s/he has to reinterpret the utterance. However, if the hearer simply does not get it and regards the sentence as incomplete or ungrammatical, this means that the (temporary) ambiguity is not perceived.¹⁸⁷

(45) The old man the boat.

It has to be admitted that the example cited in (45) is to a certain degree artificial, as it has been constructed in linguistic research in order to illustrate similar garden path effects. However, we will see that similar cases of ambiguities are also attested in everyday language (cf. examples (58) and (59) below). Still another example of ambiguity can be found in (46):

(46)

In this headline we are dealing with a wordplay, that is, a deliberate game between the speaker / writer and the hearer / reader, as determined by the context of the utterance (the publication in a satirical newspaper). Therefore, the hearer can be expected to be suspicious, looking for possible cases of ambiguity and wordplay in the headlines s/he reads, and this leads us to our next parameter which analyzes the strategic vs. non-strategic use of ambiguity by the communication partners:

The strategic use of ambiguity can be either acknowledged by the hearer or ignored by him/her – at the same time, an ambiguity may also be used by a speaker non-strategically, which means unintentionally or by mistake, while it is, again, either ignored by the hearer or recognized and then understood in a manner different than the one intended by the speaker. The latter is the case in (47): here, the speaker, a company wanting to sell electrical domestic appliances, certainly did not use the ambiguity strategically.

(47)

In this advertisement, which shows a woman doing the hovering, the reference of the personal pronoun is unintentionally ambiguous: “it” can either refer to the NP “hard work” or to the proposition as a whole (see also Wales 1996, 49).

(48) Don’t let hard work kill your wife, let electricity do [the work].

(49) Don’t let hard work kill your wife, let electricity do [the killing of your wife].

Whereas the intended meaning of the advertisement is to promote the use of electrical equipment in the household, the slogan can also be read differently. Given the genre of advertisement, we can only assume that this meaning is unintended by the speaker (especially so as the joke in reading (49) does not make any sense in this kind of advertisement). The hearer / reader, however, may even more immediately choose reading (49) and not understand “it” as referring to the NP.

A similar problem can be found in the following newspaper headline:

(50) Squad helps dog bite victim (Cooper 1980)

Again, two different readings are possible, although only one is probable (as intended by the author). The probable (intended) reading is of dog bite victim as a compound noun, with the head victim being specified by the modifier dog bite (with a hyphen missing); the other grammatically possible reading is dog and victim as two separate nouns and bite as a verb.

(51) squad helps _N[dog bite victim]

(52) squad helps _N[dog] _V[bite] _N[victim]

While examples (47) and (50) have turned out to be unintentionally ambiguous and comical, authors in newspaper headlines may also want to be intentionally so, as in (53), a French newspaper headline from 1970, when Michel Debré was state secretary of education and denied that aircraft deliveries to Libya had taken place.

(53)

Here, an ambiguity arises, as dément can be interpreted either as a verb form (F. démentir ‘deny’) or as an adjective meaning ‘insane, mad’. As this headline appeared in the satirical newspaper Canard enchaîné, which is characterized by frequent use of wordplay in its headlines, the reader already expects strategic ambiguities. Moreover, in this case, the ‘insane’-reading can even be supposed to be the intended one, as there was no article which followed the headline. We are thus dealing with a context that determines an intentional play with linguistic ambiguities between the speaker(s) and the hearers / readers.

This kind of strategic use of ambiguity is also often used in literary texts, for instance in chapter headings or titles of whole works:

(54)

(55) Scales of Justice (Marsh 1956)

The chapter title from Dickens plays on the double meaning of chapter. This word refers to the members of the administrative body of the church – which chimes in with the meaning of dean as the person who is in charge of church property and fits with the content, as it is about the conversation between the dean of a church and his verger.¹⁸⁸ But the word chapter can also be understood in terms of a self-referential hint to the chapter of the book that is being read.

Things are more complex in the title of Ngaio Marsh’s detective novel. The primary interpretation of Scales of Justice is to read it as a synonym of the balance that Justice holds, again in direct connection with the genre: in a crime novel, it is more than likely that justice is finally attained. Towards the end, however, the reader becomes aware of a second meaning: the scales of a fish lead detective Inspector Alleyn on the right track and help him find the murderer, the scales thereby become ambiguous in discourse and no longer refer to a balance only. Whereas, at the beginning, the word is unequivocal (at the level of discourse), in the process of reading, it attains a second discourse meaning which is determined by the plot of the novel (and which relies, of course, on the ambiguity of scales on the level of the linguistic system). Ngaio Marsh here creatively coins an ambiguous expression by actualizing a second meaning within the collocation which is already contained in the language, namely on the level of the linguistic system.

Another parameter of analysis which can be applied to the study of ambiguity in discourse is the question whether an ambiguity remains non-resolved:

For the latter alternative, we will distinguish between three basic types of factors that may contribute to disambiguation: time, i.e. cases in which the ambiguity is only temporary and disappears during the processing of the utterance, context, and, thirdly, the use of metalinguistic strategies by the communication partners in order to overcome the problems arising from an ambiguity. These factors may work differently on the various communicative levels within a literary text: Ambiguity may both arise and disappear through the interplay of those levels, and it may arise and disappear on the various levels involved in a different manner.

However, before analyzing the various possibilities of successful disambiguation, let us start by having a look at an example where ambiguity leads to a misunderstanding and to miscommunication, which means that a disambiguation would be necessary, but does not occur or, rather, fails.

(56)

The ambiguity is here based on homophony:¹⁸⁹ The White Queen perpetually misunderstands Alice as she misinterprets the homophones flour / flower, ground / ground. The fun is less on the characters’ than on the reader’s part in this case, as the reader is faced with the Queen’s message in a different realization: s/he immediately recognizes these puns as s/he sees the words written on the page.

The comedy in this example is based on the fact that Carroll not only plays with homophones but also with a total negligence of context (which should permit a disambiguation here) on behalf of the hearer: it should actually be clear to the White Queen that, in the context of baking bread, Alice does not mean blooming flowers but flour as this is needed for making bread. At the same time, Alice’s reaction can be seen as a sort of metalinguistic strategy: she corrects the White Queen (“it isn’t picked … – it’s ground”) in order to elucidate that there has been a misunderstanding of flour, but the ambiguity still remains, that is, the White Queen does not correctly identify the intended meaning of flower, but runs into a further misunderstanding (of ground) instead.

The play we encounter here is one based on the interpretation of homophones in the alternative sense, which then is the basis for the next question Alice is being asked: bread – flour > flower – pick > pick – ground > ground – acres. The result is that any successful communication collapses as speaker and hearer continue to talk of different things – although from a purely phonetic viewpoint, their interpretations are both possible.

Let us now focus on some examples of successful disambiguation. As we have seen, a first way in which ambiguities may be resolved is the continuation of the utterance itself, that is, ambiguities may be temporary and disappear in language processing.

An interesting case in point in this context is the beginning of Charles Dickens’s A Tale of Two Cities (see Bauer & Zirker 2014):

(57)

The views presented here are mutually exclusive and going to extremes, by which the contradiction is enhanced (cf. Bauer & Zirker 2014). The contradictory statements found in the first paragraph of the novel are apparently the narrator’s own, which means that he either is unreliable or that we are to “reinterpret his contradictions as ways of saying that the time (age, epoch etc.) itself was so extreme and contradictory that there is no other way of describing it” (Bauer & Zirker 2014, 213). These possible meanings are, however, reversed again after the dash, and thus the ambiguities mentioned so far – on the narrator’s part or on the level of the subject matter itself – “become part of another ambiguity” (Bauer & Zirker 2014, 213): now the statements of the first part of the sentence are revealed to be those of the “noisiest authorities”. What the reader has so far taken to be descriptions of the time are actually the opinions of an unlikable set of authorities: In a series of statements about the time and age, Dickens ironically plays with the reading time, i.e. with the readers’ progress towards a point at which they reflect upon and reevaluate what they have just read (see Bauer & Zirker 2014); this means that here both the progression and the mental processing are involved. Through the process of reading the ambiguity established at the beginning of the paragraph is thus resolved by its end; the ambiguity is only temporary (cf. section 3.1 and Wasow, this volume).

(58)

(59)

The advertisement by the German tabloid newspaper BILD aims at a similar effect, which is underlined by the typography chosen. At first, the message in (58) reads Bild stops – but then it is reinterpreted to Bild obeys its readers when we go on reading the small letters. (59) works in a similar manner: the large letters convey the message Bild gives up, which steers the reader’s attention towards the message, to then redirect the meaning towards a different one (Bild accelerates on every page).¹⁹⁰ Here again, it can be assumed that the (temporary) ambiguity is used by the speaker in a fully deliberate way (being guided by marketing strategies), and that it is being perceived by the hearer / reader.

In popular culture, a very similar mechanism of reinterpretation is used in so-called NOT-jokes:

(60)

(61)

In these examples, an utterance is followed by its negation by means of adding “NOT” at the end. The structure of this kind of utterance can be compared to (58) and (59): the meaning is turned into the opposite by concluding the utterance with the negation particle “not” in order to make a joke or to qualify (rectify) an otherwise wrong or potentially risky or even appalling statement. The ambiguity is a temporary one as the hearer cannot be sure if the statement has been brought to a conclusion by the speaker or not. The speaker’s intention is to provoke a reaction from the hearer while his utterance is still unfinished. This reaction then proves ridiculous or inappropriate as soon as the speaker ends his sentence by adding NOT and thus establishes a meaning contrary to the one of the unfinished sentence.¹⁹²

In other kinds of utterances, however, no similar effects of a forced reinterpretation by the hearer in the course of language processing occur. Nevertheless, many language-based ambiguities are resolved by other means, e.g. by contextual factors that determine the (most plausible) interpretation of the utterance, i.e., potentially ambiguous utterances are interpreted as unambiguous by the (cooperative) hearer.¹⁹³ This permits us to formulate another key parameter of analyzing ambiguity:

We have already seen some examples illustrating similar processes of disambiguation of potentially ambiguous signs in sections 2 and 4.1:

(62)

(63)

As we have seen, in each case one of the meanings of bank is immediately ruled out by the context. This latter term is understood in a broad sense here, including both the linguistic and the extra-linguistic context; at the same time, we can read all these different types of contexts as elements of the speaker’s and hearer’s knowledge, including knowledge of the language system and the conventional meanings of the words, knowledge of frequent collocates, cultural knowledge etc. (a comprehensive overview and discussion of context types is given by Coseriu (1955–56) and Aschenberg (1999); for some more reflections about different types of contexts and their role in disambiguation cf. Winter-Froemel, 2013, 156). In (62), we can find the word money in the linguistic context which also evokes a certain extra-linguistic context, and it can be assumed that a conceptual frame is activated which contains elements such as MONEY, SAVE and FINANCIAL INSTITUTION, etc. In (63), disambiguation is triggered by the mentioning of river etc. and by the extra-linguistic context and world knowledge (e.g. the possibility of constructing a road on a bank); moreover, the collocations of bank (“on a”, “of the river”) play a central role here.¹⁹⁶ Moreover the linguistic context in both cases differs with regard to the preposition used (in a bank vs. on a bank), which equally contributes to the disambiguation of the word bank.

Interestingly, the following example (which has already been discussed above) shows that speakers (e.g. literary authors) can also manipulate the hearer’s / reader’s interpretation by introducing a new context of interpretation: as we have seen, (64) is at its primary reading most probably interpreted as unambiguous (given the context of justice and the conventions of the text type of a crime novel), but the new context introduced at the end of the novel reactivates the other meaning of scales (the scales of a fish), thereby leading to the ambiguity of a heretofore unambiguous term.

(64) Scales of Justice (Marsh 1956)

Let us finally note that there are also cases in which hearers are not cooperative, that is, the context available should actually permit them to disambiguate the message, but they deliberately choose to insist on the ambiguous or on the alternative (implausible) interpretation. Cases like this frequently occur in classrooms with pupils in their puberty.

(65)

In ordinary circumstances, the teacher’s utterance would not be interpreted as ambiguous: given the extra-linguistic context – a lesson in physics – the interpretation of Gummi should be clear (we can assume that there is a concrete referent –a rubber strap forming part of a technical device – which the teacher refers to by using the noun with the definite article), and the teacher now expresses his fear that the rubber strap could be torn. However, similar utterances by teachers frequently provoke laughter, as pupils detect a further, sexual interpretation: as G. Gummi also has the meaning ‘preservative’ (English equivalent translation rubber) on the level of the language system, the sentence can also be interpreted as expressing the fear of a preservative being torn. In this case, the hearers thus choose to be uncooperative – they do not try to interpret the message in the sense intended by the speaker, but they follow a different strategy of maximizing ambiguities comprising (mainly) sexual interpretations.

Another way of achieving disambiguation is the use of metalinguistic strategies by the speaker and hearer:

In some cases, speakers are conscious of the potential ambiguities of the signs used, and a possible ambiguity of their actual utterances results from this system ambiguity. In these cases, in order to avoid misunderstandings, they may add metalinguistic comments referring to the level of orthography, as in (66) (disambiguation of sie ‘she / 3Pl’ vs. Sie ‘2Sg/Pl polite form of address’), or by specifying the intended meaning by adding a further semantic specification, as in (67) (remarkably, both forms also appear in the OED, where they are marked as colloquial forms used to distinguish both senses of funny).

(66) G. Sie groß geschrieben (‘Sie’ with capital letters, translation EWF/AZ)

(67)

Whereas (66) and (67) are examples of the attempt to avoid ambiguity and a misunderstanding resulting from it – the meta-comment in the utterance has exactly this function –, (68) illustrates an exchange between two characters in a novel that is based on a misunderstanding:

(68)

Reverend Helstone means a different kind of arms than his curate Malone as he has a particular context in mind, namely that of weapons and of defending one of the neighbouring mills against revolting workers. This context is either not known to Malone or intentionally ignored by him, and he therefore interprets “arms” differently. This exchange has the function of comic relief for the reader to whom the context weapons was introduced earlier in the chapter – “requesting to be informed whether it was the mode in Ireland for clergymen to carry loaded pistols in their pockets” (Brontë 2008, 9) – and is now taken up again with reference to Malone, who is an Irishman. In a highly overwrought situation, this tension is slightly resolved through laughter evoked on the level of author-reader-communication ¹⁹⁸; it is not meant as such for the characters within the novel. At the same time, the metalinguistic comment “weapons I mean” is a means of remedy to allow the communication to go on despite the initial misunderstanding.

4.3 Ambiguity and Levels of Communication

We have seen that in the conversation between Alice and the White Queen in (56) several levels of communication are involved: the one between the characters in the text as well as the one between the author and reader at whom the joke in the exchange is actually directed. This use of ambiguity can be described as a typical feature of literary texts (cf. above 3.2), and it leads us to another key parameter of analyzing ambiguity:

The multiplication of communicative levels becomes even more pertinent and immediate in the theatre when the audience is included in a joke by the author and the joke is not at all directed at the characters involved in the scene. The following example from Shakespeare’s Romeo and Juliet is such a case. Mercutio, Romeo’s friend, has just been mortally wounded by Tybalt:

Mercutio here makes a joke based on the ambiguity of “grave”, referring both to SERIOUS as well as to TOMB, which alludes to the fact that he will be dead soon and in the grave. Throughout the play he has been making jokes, but still the audience ask themselves why he would try to be funny at this very moment when it appears to be quite inappropriate.¹⁹⁹ At closer inspection, however, the joke seems less directed at the other characters present on the scene, especially so as Mercutio also cries out “A plague o’ both your houses” several times (Shakespeare 2006b, 3.1.92; 100–101; 108), but rather at the audience for whom it serves as an instant of comic relief in an otherwise very serious (i.e. grave) moment. In addition to the speakers and hearers onstage, another level of communication is revealed here, namely between the author – who created the joke – and the audience.²⁰⁰

This doubling is not only made use of in all kinds of literary texts (see, e.g., (23), (56), (68) and (69)) but also in other media, for instance, in advertisements and movies. Moreover, it can be linked to the concept of polyphony. Originally a term from musicology, this concept has been transferred to the domain of literary studies, where it has been coined by Bakhtine’s approach towards the dialogicity of literary texts (Bakhtine 1979 [1934/1935], 1984). Then, Ducrot applied this concept to linguistic investigations, focussing mainly on the diversity of voices at the speaker’s side and distinguishing various instances of speaker roles (locuteur L and λ, sujet parlant, énonciateur, cf. Ducrot 1984). Taking up this line of reasoning, pragmatic analyses of phenomena such as irony have been recently proposed which consider the multiplication of points of view to be a central feature of the phenomena studied (Rabatel 2006, 2008, 2012). Here again, by focusing on the communication process as a complex interaction of various speaker groups and various roles of speakers (and hearers), very fine-grained analyses of individual manifestations of speaker-hearer interaction can be obtained.

4.4 Ambiguity in a Diachronic Perspective

Most of the examples analyzed so far have illustrated cases in which the ambiguity on the discourse level results from the signs (or syntactic structures) being already ambiguous on the level of the language system, that is system ambiguities are actualized in discourse (cf. examples (68) and (69) above). However, we have also encountered some cases in which a new interpretation (beyond the lexicalized meaning of the word) is introduced into the communication process (cf. examples (11), (12), (35) and (36)). This leads us to a parameter which is of central interest for linguistic studies focusing on processes of language change, and more specifically semantic change, as it takes up the question how new meanings can enter the language system.

The latter kind of ambiguity represents a very interesting case which has not been taken into account in traditional linguistic approaches to ambiguity. It is strongly linked to language change: In a general manner the initial stage of processes of semantic change is characterized by the introduction of a new (additional) meaning besides the conventional meaning, which occurs in certain contexts. This new meaning may then diffuse among the speech community, being reused by other speakers and in other contexts, ultimately leading to a change in the linguistic convention and, in some cases, to a loss of the original meaning (cf. Heine 2003, Traugott & Dasher 2002, 12, Kuteva & Heine 2008, 224). Yet the traditional formula describing semantic change A → A/B → B is not entirely clear, as can be shown by referring to the distinction between ambiguity in the language system and ambiguity in discourse introduced in section 2. In traditional approaches, A and B are not only interpreted as entities in the language system (i.e. the conventional meaning of a given item), but also as entities on the discourse level. However, this means that two different aspects are mixed up here, and the intermediate stage is sometimes interpreted in terms of a situation of polysemy (A/B = ambiguity in the language system; cf. Bréal 1897), but also referred to as concrete situations of communication where the item is ambiguously used (A/B = ambiguity in discourse). This latter aspect, which has also been discussed with regard to the concept of bridging contexts introduced by Evans & Wilkins (2000, 549–550, cf. also Heine 2002, 85–86, Marchello-Nizia 2006, 253–264, Benveniste 1966, 290 and note 33 above) seems particularly important for our concerns, as it marks the moment of innovation proper (the first introduction of the new meaning in discourse) which precedes the stage of polysemy.

It thus seems very promising to relate ambiguity to the field of language change by analyzing a few examples.²⁰¹

(70)

(71) F. C’est pas terrible. (That is not great / mediocre, translation EWF/AZ)

In utterances like (70), the French word terrible is interpreted in a sense markedly different from its original meaning (‘terrible’), and it assumes a clearly positive value (‘ingenious, brilliant’); this new interpretation can also be found in the frequently used expression cited in (71).²⁰² How can these uses be explained? The two meanings stand in a relation of conceptual contrast, and it can thus be assumed that the novel use occurs in contexts marked by the speaker’s intention to be maximally expressive and to call the hearer’s attention by his “striking” message, which forces the hearer to adapt his interpretation of terrible to the context. In Gricean terms, this process could be seen as an apparent violation of the maxim of quality (cf. uses of irony in general), which triggers a conversational implicature towards the novel interpretation (cf. Grice 1975). Let us note that this use has already been conventionalized, i.e., the word nowadays has both meanings and is thus ambiguous on the level of the language system as well (cf. DHLF s.v. terrible).²⁰³ So contexts of expressivity and ironical contexts represent the (potential) starting point of processes of change leading to a conventionalization of the new interpretation (which implies that the expressive or rhetorical effect is diminished, as can be seen in example (71)).²⁰⁴

In other cases, in contrast, the introduction of a new interpretation can occur without such effects of expressivity. Consider the example of the English word scales, which can designate the two pans on a balance, but also the balance as a whole. Here we can assume that the semantic change has originated in contexts where both interpretations are fully compatible with the interpretation intended by the speaker.

(72) scales ‘the two pans on a balance’ / ‘a balance’ (OED)

Another interesting case of language change is illustrated in (73): the original meaning of the German word Traube is ‘bunch of grapes’, but today it is also frequently used in the sense of a single ‘grape’ as well (replacing the older form Weinbeere).²⁰⁵ Here again, in certain contexts both interpretations are possible, so that a situation of (potential) ambiguity arises – the hearer may interpret the word in its novel sense without any negative effects on ongoing communication (cf. (74); by contrast, in other contexts of use the interpretation of the form seems to be biased towards one of the two meanings, cf. examples (75) and (76)).²⁰⁶

(73) G. Traube ‘bunch of grapes’ / ‘grape’

(74) G. Trauben zu verkaufen (‘bunches of grapes / grapes for sale’)

(75) G. Trauben ernten (‘to harvest bunches of grapes’)

(76) G. Trauben essen (‘to eat grapes’)

The ambiguity is here located on the conceptual level; that is, in the contexts of use where an ambiguity turns up, the referent of Trauben can be conceptualized both as BUNCHES OF GRAPES or simply as GRAPES, without any problem arising in an ongoing communication.

As we have seen, the examples cited so far in this section share the basic characteristic that they represent cases of new ambiguities that are introduced on the level of discourse, that is, signs that are not ambiguous on the level of the language system are used in a way which permits or requires an interpretation in a new sense; this new sense is, at a first stage, purely pragmatic, but may become conventionalized at a later stage of time.

Despite this common feature, the examples also show important dissimilarities with respect to the way in which ambiguities arise in communication. A basic criterion which can be used to distinguish two main cases depends on the communication partner who introduces an ambiguity:

As we have seen, examples like (70) can be explained by referring to certain strategies the speaker pursues in communication; another example similar in this respect is given in (77) (for a more detailed discussion see section 2). In all these cases, we are thus dealing with speaker-induced change, or a speaker-induced introduction of ambiguity.

(77) While you were lying in the sun, I washed the dishes.

In other examples, in contrast, it is the hearer who plays the crucial part (cf. example (74) above), and these cases can be treated as instantiations of hearer-induced change, or of a hearer-induced introduction of ambiguity. Recent research has shown that similar processes of hearer-induced semantic reinterpretation represent an interesting subtype of language change, which is studied under the label of reanalysis (cf. Detges & Waltereit 2002).

However, if we deal with ambiguity in speaker-hearer-interaction, we must bear in mind that the simple distinction between speaker- and hearer-induced ambiguities is, to a certain degree, oversimplified, as we can always presume an interaction of the speaker and the hearer in communication. For example, as we have seen, speaker-induced changes such as regarding while presuppose that the hearer ratifies the novel use, and conversely the explanation of hearer-induced change implies in a trivial way that the speaker’s utterance (and the speaker’s meaning) is taken into account as the basis on which the reinterpretation occurs.

Moreover, certain examples of ambiguity reveal a further complexity arising from the fact that different speakers and / or hearers are “available”, that is, in some cases communication is located on various levels simultaneously (cf. section 4.3).