2.1 Givenness

Existing studies agree that sentence-internal jo encodes that the proposition is ‘given’ or ‘uncontroversial’ for the addressee in some sense or other. Solberg (Reference Solberg1990) concludes that jo marks what is said as known or given. Using Grice’s terminology, Andvik (Reference Andvik1992:85) states that jo conventionally implicates that there is consensus about the facts conveyed by the jo-clause. Lind (Reference Lind1994:104–125) uses the term ‘interpersonal’ and writes that jo creates a shared context where the hearer is assumed to share the speaker’s knowledge and assumptions. Fretheim (Reference Fretheim1991) states that jo expresses that the speaker considers the truth of the expressed proposition p to be ‘mutually known’ to the interlocutors. Mutually known includes situations where the speaker thinks the addressee ought to know, or should be able to infer, that p is true (Fretheim Reference Fretheim1991:184).

A typical use of jo is illustrated in (2) below. The excerpt is from the BigBrother Corpus. The conversation takes place in a reality show on TV where the participants share a house (see also Section 4.1). Anette, Lars, and Rodney are all in the bathroom. Anette is about to take a shower, and Lars teasingly suggests that he and Rodney stay in the bathroom. Rodney points out, using an utterance with jo, that he has seen Anette naked ten times.Footnote ⁴

1. (2)

   

   (BigBrother Corpus)

Rodney and Anette both know that Rodney has seen Anette naked on several occasions, which licenses his use of jo. If Rodney wanted to inform Anette about a fact that would be surprising to her, sentence-internal jo would not have been natural.

The previous descriptions of jo’s givenness aspect are consistent with cases like (2), but it is possible and desirable to sharpen the analysis. As mentioned above, Fretheim (Reference Fretheim1991) states that jo communicates mutual knowledge, and additionally covers cases where the addressee ought to know the fact described in the sentence or should be able to infer it. This disjunctive description is not wrong, but a unitary and explanatory analysis would be preferable. Lind’s (Reference Lind1994) description of jo as an expression which creates a shared context – and her demonstration of the various ways jo is used in communication – is also not wrong. It is though slightly vague, and it is not a semantic proposal or an account of the pragmatic processes involved in the comprehension of utterances with jo. As for Andvik’s account, his use of ‘consensus about the facts’ in his description of the basic meaning of jo (Andvik Reference Andvik1992:86) makes some wrong predictions: people can, as we shall see in Section 4, use jo when presenting something the addressee does not agree on.

We propose that the givenness aspect of jo is best understood as mutual manifestness (Sperber & Wilson Reference Sperber and Wilson1986/1995:43–44; Section 3.4 in this paper). That is, our semantic proposal includes a constraint such that the utterance’s proposition is to be entertained as mutually manifest to the interlocutors. As we shall demonstrate in Section 4.2, mutual manifestness captures the ‘givenness’ aspect of interpretations of utterances with jo in authentic communication. Moreover, the notion of mutual manifestness has been shown to be important on independent grounds (Sperber & Wilson Reference Sperber and Wilson1986/1995) and should hence be favoured over competing analyses, other things being equal.

2.2 Premise for deriving a conclusion

The second issue is whether or not sentence-internal jo signals that the proposition expressed in the utterance ought to be interpreted as an argument for a further inference. Fretheim (Reference Fretheim1991) does not mention this meaning aspect in his account of jo. Solberg (Reference Solberg1990:66–68) demonstrates that jo can be used to mark the information as a premise for an inference, but nevertheless concludes (Reference Solberg1990:76) that the ‘basic meaning’ of jo is merely to mark the information as known or given. Other functions, like marking the information as a premise, are analysed as uses that derive from this basic meaning.

Andvik (Reference Andvik1992) convincingly argues that consensus about the propositional content is not enough to license jo in an utterance. For instance, jo is not appropriate in answers to quiz-questions, even though the information in the answer is known to the hearer. This is illustrated by the invented example in (3):

1. (3)

   

   (adapted from Andvik Reference Andvik1992:60)

The pupil Geir cannot use jo in (3) – at least not if he merely intends to supply the information represented by the proposition and thereby answer the teacher’s questions. However, if Geir offers the information in the utterance as an argument for some conclusion he is drawing, he can use jo (Andvik Reference Andvik1992:61). Interestingly, this meaning aspect is not directly reflected in Andvik’s final analysis: Andvik (Reference Andvik1992:85–86) concludes that jo has two meaning aspects: (i) ‘there is consensus about the facts’ and (ii) ‘there is an idea in the air which will be countered by this consensus’.

Andvik’s observation concerning the invented example in (3) is supported by the authentic example rendered in (2) above. Rodney uses jo in the utterance where he claims that he has seen Anette naked several times, and the presence of jo makes it clear that Anette should draw further conclusions from this information. Since Rodney and Lars are present in the bathroom even though Anette is about to take a shower, a natural interpretation is that Rodney’s utterance provides an argument for staying in the bathroom. A possible line of inference is explicated in (4):

1. (4)

   

Without jo, the argumentative purpose of Rodney’s utterance in (2) – i.e. the intention to argue in favour of the conclusion rendered in (4) – would not have been as clear. Without jo, the utterance could have been used to introduce a new topic whereas with jo, this possibility is ruled out. The presence of jo in (2) signals that the utterance’s content is to be taken as a premise which relates to an issue under discussion. This supports the following proposal which will be tested in Section 4.3: The semantics of jo includes a constraint such that the utterance’s proposition is interpreted as a premise for deriving a conclusion.

2.3 Opposition

Fretheim (Reference Fretheim1991) does not mention any oppositional aspect of sentence-internal jo at all, and according to Solberg (Reference Solberg1990), rhetoric purposes is just one of the functions that jo can serve. Andvik, on the other hand, argues that an utterance with jo is used to counter or oppose some actual or fictive idea assumed to be ‘in the air’ (Reference Andvik1992:75–86). Lind (Reference Lind1994) agrees with Andvik that jo is oppositional in the sense just described, and (2) above seems to support their view, in that the conclusion in (4) opposes Anette’s requirement that Rodney and Lars should leave the bathroom. However, jo may also be felicitously used in cases without apparent opposition. Consider the following example:

1. (5)

   

   (Oslo Corpus)

In (5), the utterance with jo backwards supports the content of the previous utterance: since the fish hall in the Ravnkloa area is so ugly that it is a shame for the town, it should be demolished. The question is whether the conclusion derived from the jo-utterance – i.e. that the fish hall should be demolished – opposes an idea. There is no explicit contrasting view to oppose in the context in which (5) occurs, but one could of course imagine that someone would hold such view – after all, the fish hall has not yet been demolished, which suggests that at least some people want to keep it. Andvik (Reference Andvik1992) argues that such cases of ‘seemingly supportive jo’ still involve opposition, and that the opposition is against a hypothetical objection.

We agree with Andvik (Reference Andvik1992) and Lind (Reference Lind1994) on the point that jo is often used in contexts where the speaker argues against someone’s view. As we shall see in Section 4.4, however, our data set does not support the hypothesis that a jo-warranted conclusion always counters an idea in the air. We therefore propose that the semantics of jo includes a constraint on the interpretation such that the propositional content is interpreted as a premise for drawing a conclusion (in addition to the constraint on mutual manifestness). We shall in due course demonstrate how the two proposed semantic constraints predict interpretations of utterances with jo in authentic discourse. But first, a presentation of the necessary theoretical concepts is in order, as these are crucial to a precise account of which (semantic) meaning jo brings to the utterance interpretation, and how and why this meaning sometimes – but not always – gives rise to context specific interpretations such as opposition.

3.1 General principles

By semantics we mean conventionally encoded meaning in the sense of Blakemore (Reference Blakemore1987), Carston (Reference Carston2002, Reference Carston2008) and Ariel (Reference Ariel2008, Reference Ariel2010),Footnote ⁵ that is, linguistic meaning which is part of the language code rather than the result of pragmatic inference (see also Bach Reference Bach2001). Thus, the semantics of jo is the lexical knowledge which is activated when the expression is used in an utterance, feeds into the interpretation process, and interacts with contextual properties and pragmatic processes in yielding the various context specific interpretations (see Sperber & Wilson Reference Sperber and Wilson1986/1995:172–173; Carston Reference Carston2002:22, Reference Carston2006, Reference Carston2008). The interpretation of a pragmatic particle is highly context dependent. The study of pragmatic particles is therefore likely to benefit from a holistic pragmatic theory which observes the interaction between linguistic semantics and utterance interpretation. Relevance Theory (Sperber & Wilson Reference Sperber and Wilson1986/1995, Wilson & Sperber Reference Wilson, Sperber, R. Horn and Ward2004) is a cognitive pragmatics framework with a more than 30 years old tradition for dealing with non-truth-conditional semantics, which was initiated by Diane Blakemore’s work on procedural meaning (Blakemore Reference Blakemore1987, Reference Blakemore2002; see also Wilson & Sperber Reference Wilson and Sperber1993).

An important aspect of Relevance Theory (Sperber & Wilson Reference Sperber and Wilson1986/1995) is that it recognises that human beings are relevance-oriented. The notion of relevance is a technical one: relevance is a measure of efficiency in cognitive processing of stimuli, including utterances. All else being equal, a stimulus is relevant to the extent that it yields positive cognitive effects (i.e. improves the individual’s mental representation of the world by confirming or eliminating the individual’s existing assumptions or by yielding new conclusions) and to the extent that the processing effort involved in computing these effects is relatively low. According to the communicative principle of relevance, audiences are geared towards identifying the optimally relevant interpretation of an utterance (Sperber & Wilson Reference Sperber and Wilson1986/1995). This means that when faced with a semantically underspecified utterance – as utterances are (Carston Reference Carston2002) – the audience is geared towards the interpretation that yields sufficiently many positive cognitive effects for no unjustifiable processing effort (Sperber & Wilson Reference Sperber and Wilson1986/1995:270). The search for optimal relevance – i.e. a balance of processing effort with positive cognitive effect – governs the selection of contextual assumptions as well as pragmatic processes such as disambiguation and the derivation of implicatures. In short: utterance interpretations are constrained by the semantics of the linguistic expressions in it, cognitive pragmatic principles of relevance, and contextually available assumptions.

3.2 Explicatures and implicatures

The sub-tasks involved in the overall comprehension process includes the construction of a ‘hypothesis about explicit content (explicatures) via decoding, disambiguation, reference resolution, and other pragmatic enrichment processes’ (Wilson & Sperber Reference Wilson, Sperber, R. Horn and Ward2004:615). In the case of ordinary assertions of declarative sentences, the result is a truth-evaluable description. Given that this proposition is intended to be communicated as a description of a state of affairs in the world, it constitutes the utterance’s basic explicature. Expressions which affect the truth-conditions of an utterance contribute to the utterance’s basic explicature (Wilson & Sperber Reference Wilson and Sperber1993). Sentence-internal jo never affects the truth-conditional content in an utterance. Rodney’s utterance in (2) above, for instance, is true and false in exactly the same situations regardless of whether jo is present or not.

Languages not only allow speakers to describe states of affairs in the world; sometimes the most relevant interpretation of an utterance is which attitude the speaker expresses towards it (e.g. epistemic certainty) or how it relates to other contextually available propositions (e.g. opposition or support). A communicated assumption that embeds the propositional content of the utterance under such higher-order description is called a higher-level explicature (see e.g. Carston Reference Carston2002:377). Languages display different resources – such as pragmatic particles in the case of Norwegian and some other Germanic languages – that are used to guide the derivation of higher-order explicatures rather than contributing to the basic propositional content. The constraint on mutual manifestness we propose for jo is a constraint on the derivation of a higher-level explicature such that the propositional content is entertained as mutually manifest (see also Blass Reference Blass, Andersen and Fretheim2000).

Another central category of communicated meaning is implicature, a category that was originally introduced by the philosopher Paul Grice (see Grice Reference Grice, Cole and Morgan1975). Within the Relevance-theoretic framework, implicatures are defined as ostensively communicated assumptions that are not explicatures; thus they are derived solely through inferences (Carston Reference Carston2002:377).Footnote ⁶ The input to the inferential process of deriving implicatures are the explicature(s) and contextual assumptions, and the inferential process is guided by the principles of relevance.Footnote ⁷ The Relevance-theoretic implicature corresponds roughly to the Gricean conversational implicature. The Gricean conventional implicature, on the other hand, is not recognised within Relevance Theory, because it is a conventional link between a form and a meaning. As such, it falls under encoded meaning and hence it is a semantic phenomenon. The term implicature in the Relevance-theoretic sense is reserved for communicated meaning that results from pragmatic inference based on the utterance’s explicature plus contextual assumptions. This means that a given linguistic expression does not encode a certain implicature, but it can encode constraints that guide the process of deriving one. The proposed constraint on the proposition as a premise for a conclusion is a constraint on the derivation of implicatures, because jo warrants this derivation of a conclusion. Note that implicatures are usually accessible without any linguistic expression that constrains their derivation. However, as we shall argue in Section 4.2, jo makes the route to the implicature more accessible and thereby arguably decreases the processing effort spent on achieving cognitive effects.

3.3 Procedural meaning

Relevance Theory distinguishes between linguistic expressions that encode procedural meaning (e.g. but, so, and after all) and linguistic expressions that encode conceptual meaning (e.g. dog, run, and happy) (Blakemore Reference Blakemore1987, Reference Blakemore2002; Wilson & Sperber Reference Wilson and Sperber1993; Wilson Reference Wilson, Escandell-Vidal, Leonetti and Ahern2011). Whereas conceptual meaning is relatively concrete and can be brought to consciousness, procedural meaning is vague and hard to paraphrase. Procedural meaning is a semantic phenomenon just like conceptual meaning, because both types of meaning are encoded constraints. The difference is that conceptually encoded meaning supplies constrains on which ad hoc conceptual representation to construct, whereas encoded procedures ‘constrain and guide pragmatic processes which are essential in deriving the intended interpretation’ (Carston Reference Carston2016:159; see also Blakemore Reference Blakemore1987, Reference Blakemore2002; Wilson Reference Wilson2016). Sentence-internal jo clearly belongs to the group of procedural expressions. That is, the proposed constraint on mutual manifestness is a procedural instruction to construct a higher-level explicature and look for cognitive effects based on the propositional content as mutually manifest. And the proposed constraint on implicatures is a procedural instruction to use the propositional content as a premise for deriving an implicature.Footnote ⁸

Wilson (Reference Wilson, Escandell-Vidal, Leonetti and Ahern2011) has suggested that procedural meaning can be viewed as linked to sub-heuristics of general cognitive capacities such as mindreading, emotion reading, social cognition and argumentation (see also Sperber Reference Sperber, Carruthers, Laurence and Stich2005). According to Wilson (Reference Wilson, Escandell-Vidal, Leonetti and Ahern2011), one prediction from this view is that we might expect to find clusters of procedural items linked to such domain-specific capacities. In Section 5.1, we shall argue that the procedures encoded by jo activate the capacities for epistemic vigilance and evaluation of arguments. As we shall see, this explains the distribution of opposition interpretations of utterances with jo.

3.4 Mutual manifestness

Pragmatic inference not only rests on context in terms of assumptions that are known to the speaker and addressee. The construction of context also exploits assumptions that are manifest to them both (Sperber & Wilson Reference Sperber and Wilson1986/1995:38–46). Any assumption that an individual is capable of representing mentally and accepting as true or probably true, is manifest to him. An assumption is mutually manifest to two individuals if (i) they are both capable of accessing this assumption through memory, perception or inference, and (ii) they are both aware of this mutual access (Sperber & Wilson Reference Sperber and Wilson1986/1995:41–42; see also Unger Reference Unger2018 for a good illustration of the phenomenon). Thus, for a proposition p to be manifest to an individual, it is sufficient that the assumptions necessary for assuming that p is true are available to that individual, and it is not necessary that p is already known, assumed, or privately entertained. For instance, it is manifest to me that Julius Caesar and Noam Chomsky never had breakfast together, even if I have never considered this before, because it is a conclusion which is mentally accessible to me through the assumptions I have about Julius Caesar and Noam Chomsky (Blass Reference Blass, Andersen and Fretheim2000:44; Sperber & Wilson Reference Sperber and Wilson1986/1995). Depending on the context of communication, a certain sub-set of the speaker and addressee’s mutually manifest assumptions will be easily accessible and thereby serve as contextual assumptions in the utterance interpretation process.

Sperber & Wilson (Reference Sperber and Wilson1986/1995) use the notion mutual manifestness in their account of human communication. Later on, Blass (Reference Blass, Andersen and Fretheim2000) has shown that some linguistic expressions, such as English after all, German ja, and Hausa mana, encode constraints on mutual manifestness. While the meanings of after all, ja, and mana differ in other respects, they all instruct the hearer to entertain and store the proposition in their scope as mutually manifest. Blass analyses mutual manifestness markers as encoding procedural constraints on higher-level explicatures. An utterance of the German Er ist ja zu Hause ‘[he is at home]-ja’ can thus be used to communicate a higher-level explicature where the proposition ‘he is at home’ is embedded in a higher-order description as mutually manifest. If Norwegian sentence-internal jo is a marker of mutual manifestness, we expect it to appear in contexts similar to those described by Blass (Reference Blass, Andersen and Fretheim2000), that is, contexts where the propositional content is epistemically available to the hearer through knowledge, inference, memory or perception.

4.1 The data set

The study is based on authentic written and spoken Norwegian discourse from three searchable and digitalised corpora from The Text Laboratory (2010–2017): the Oslo Corpus, the NoTa–Oslo Corpus, and the BigBrother Corpus. The Oslo Corpus consists of newspapers, reports, law texts, and novels. For the present study we have used the sub-corpus of texts in the Norwegian written standard Bokmål. The NoTa–Oslo Corpus and the BigBrother Corpus consist of spoken discourse which is video recorded and transcribed. The NoTa–Oslo Corpus contains interviews and conversations prompted by researchers, and the BigBrother Corpus contains conversations from the Norwegian version of the reality show BigBrother from 2001. The three corpora were searched for occurrences of jo in sentence-internal position, and 174 utterances – 79 written and 95 spoken – were extracted on a random basis.

4.2 A semantic constraint on mutual manifestness

This sub-section is structured around the notion of mutual manifestness, starting with simple examples of givenness and ending with examples that could challenge our claim that jo encodes a constraint on mutual manifestness. This enables us not only to test our claim about jo’s semantics, but also to shed light on what it means for an assumption to be mutually manifest through concrete examples.

In the example in (6), jo is used in a context where evidence for the proposition is available to the interlocutors through perception. Per Morten is talking about an interaction last night between Anne Mona and himself:

1. (6) Per Morten:Footnote ⁹

2. Men i går og da så var det sånn satt vi i sofaen # så sitter og spiser og så sier a # “herregud Per Morten: har du ikke lært å spise eller” sier a # alvorlig tone egentlig da # sitter og spiser og så får jeg sånn masse smuler nedover her så hun sitter og spiser og “ja hva mener du med det” sa jeg # “ja du har jo smuler utover hele genseren” sier a.

3. (BigBrother Corpus)

4. ‘But yesterday, and it was like this we were sitting in the sofa # and sitting there eating and then she says, # “Oh my god Per Morten: haven’t you learned how to eat properly or?” she says # in a serious tone in fact # sitting and eating and then I get like a lot of crumbs down here so she is sitting and eating and “What do you mean by that?” I say # “Well, as you can see, you have crumbs all over your sweater,” she says.’

‘Well, as you can see you have crumbs all over your sweater.’

In the situation described by Per Morten, both the speaker (Anne Mona) and the hearer (himself) have access to visual evidence that Per Morten has crumbs all over his sweater, and the assumption communicated is clearly mutually manifest to them. Per Morten may already be aware of the crumbs, but in case he is not, the presence of jo in Anne Mona’s utterance helps him access the evidence and accept the assumption that he has crumbs all over his sweater. Without jo, the fact that Per Morten could observe the crumbs in the given situation would not have been as clearly communicated.

While visual evidence is one kind of situation that allows for the use of jo, it appears in our corpus data that jo is more often used in contexts where the speaker can reasonably assume that the addressee has non-visual cognitive access to the proposition expressed by the segment with jo. This tendency is seen especially in examples where the utterance’s proposition represents general knowledge that the addressee is reminded of. This is illustrated in (7) below, which is from a slightly sarcastic written review of a movie where the character Demolition Man occurs. This time, jo is part of an appositive relative clause.

1. (7) Slik skal dagens actionfilm være, opptrappet, latterlig, hysterisk. Enkelte ganger med selvironi. Og uten respekt for noe annet enn spørsmålet om hvor lydsterke eksplosjonene kan bli. Og hvor tilfeldig de kan ramme. “Demolition Man”, som jo betyr “han som ødelegger”, er en tidstypisk og på alle måter sprengfylt actionkomedie …

2. (Oslo Corpus)

3. ‘This is the way today’s action movies are supposed to get, exaggerated, ridiculous and hysterical. Sometimes self-ironic. And without respect for anything apart from the question of how loud the explosions can get. And how randomly they can target. Demolition Man, which, as you know, means “he who destroys”, is typical for its time and, by any account, a bulging action comedy …’

‘which, as you know, means “he who destroys”’

Without jo, the utterance would have been likely to offend in a country like Norway, where most adult people are expected to have some command of the English language. By adding jo, the speaker explicitly acknowledges that the information in the clause is something the addressee presumably already knows or can infer.

In (8) below, jo is used in a case where the utterance’s proposition is neither available to the hearer through perception nor retrievable from memory. In most such cases, however, it appears that the hearer should be capable of accepting the proposition as true or probably true because he has knowledge that enables him to infer that p is a true description of the world, (recall the example with Caesar and Chomsky in Section 3.4). The example in (8) below is an excerpt from Adresseavisen, which is the regional newspaper of Trøndelag County and the city of Trondheim in Norway. The person who is being interviewed has moved from her house in Parkveien, which is a nice neighbourhood, to a house close to the canal. This may appear as a strange move to some people. However, the speaker in (8) was born in an area called Ravnkloa, which is also close to the water. She therefore feels almost like coming home when moving into her new house by the canal:

1. (8) Vi har gledet oss så ustyrtelig til å flytte inn. Solgte huset i Parkveien og så frem til å bo ved kanalen. Når man er født i Ravnkloa, ja så er jo dette nesten som å komme hjem.

2. (Oslo Corpus)

3. ‘We have looked so much forward to moving in. Sold the house in Parkveien and looked forward to living by the canal. When you’re born in Ravnkloa, well, then this is obviously almost like coming home.’

‘When you’re born in Ravnkloa, well, then this is obviously almost like coming home!’

It is unlikely that the reader has ever entertained the causal relation between the two specific propositions expressed by the utterance with jo in (8). The presence of jo nevertheless gives the impression that the addressee is expected to have access to knowledge that would enable him to agree. And indeed, to people who read the newspaper, it should be common knowledge that Ravnkloa is by the water. Given the similarity between a house at Ravnkloa and a house by the canal, the claim that moving to a house by the canal is almost like coming home is mutually manifest.

Also in (9) below, the hearer is hardly expected to have entertained the specific propositional content prior to the utterance:

1. (9) Jarle Pettersen argumenterer godt for denne elektroniske varianten av dart, selv om han innrømmer at noe av sjarmen kanskje forsvinner. - Dette er en modernisering. Det er enklere, rett og slett. Og så er det jo morsomt å få en melodi når du har vunnet, sier Pettersen.

2. (Oslo Corpus)

3. ‘Jarle Pettersen argues convincingly in favour of this electronic version of darts, even though – as he admits – some of the charm is maybe lost: “This is a modernisation. It’s simpler, basically. And besides, it’s fun to get a melody when you win, of course,” says Pettersen.’

‘And besides, it’s fun to get a melody when you win, of course.’

Electronic darts is a new thing, so the addressee presumably has no experience with electronic darts. But a generalised version of the propositional content, i.e. that it is generally fulfilling to achieve a reward when one has accomplished something, is likely to be part of people’s world knowledge. The proposition expressed by the jo-utterance is hence available to the hearer through inference based on assumptions which he is expected to already possess.

There are a few instances in our data set where the evidence for the propositional content of the jo-utterance neither appears to be available through perception, nor through inference. Such an instance is shown in (10):

1. (10) Redaktøren av kronikken synes ikke helt å ha fanget Vinjes hovedbudskap. I innledningen sies det at danskenes gamle pausekommasystem erstattes av et grammatisk fundert kommasystem, mens Vinjes virkelige budskap jo er nokså nær det motsatte.

2. (Oslo Corpus)

3. ‘The editor of the article doesn’t quite seem to catch Vinje’s main message. In the introduction it is said that the Danes’ old comma system based on pauses is being substituted by a grammar-based system, while Vinje’s real message is in fact pretty much the opposite.’

‘while Vinje’s real message is in fact pretty much the opposite’

The addressee is any reader of the local newspaper Adresseavisen. It is possible that the writer in (10) assumes that the addressee has read the texts she refers to and agrees with her interpretation of Vinje’s message. But even if the writer does not assume that the addressee has read the two texts, jo is still appropriate in this context. Without jo, the writer would merely have informed the reader about p without any reference to the existence of evidence. With jo, the sentence signals that there exists sufficient evidence for the claim, and that this evidence is accessible to the reader – as well as to the editor of the article whom the writer is criticising. Since Vinje’s texts are available to the public, the proposition ‘Vinje’s message is pretty much the opposite [from what the editor assumes]’ might be seen as manifest to the reader, but only marginally so. Thus, the use of jo in (10) is pushing the limits of a mutual manifestness marker, but for a good reason: the presence of jo indicates that anyone should be able to reach the conclusion that ‘Vinje’s message is pretty much the opposite [from what the editor assumes]’ if they consult the publically available evidence. Jo thereby has the rhetoric effect of emphasising the editor’s ignorance and justifying and strengthening the writer’s criticism of him.

In the data presented so far, the speaker has no apparent reason to believe that the hearer will be reluctant to accept the propositional content of the jo-utterance as true or probably true. We now turn to cases where the utterance clearly expresses an assumption which is in conflict with the assumptions held by the hearer. In (11) below, Ramsey indirectly suggests that Anne Mona is likely to win the BigBrother prize of one million Norwegian crowns because the typical winner of such competitions is the kind and sweet little girl. Anne Mona objects to Ramsey’s world view by claiming that she is not kind at all. She uses jo in her utterance:

1. (11)

   

(BigBrother Corpus)

‘I’m not kind at all, and you know it.’

An interesting question here is whether the proposition in Anne Mona’s utterance with jo can be regarded as manifest to Ramsey, now that he has just implied the opposite view, namely that Anne Mona is sweet and kind. Whatever assumptions Ramsey has entertained up until Anne Mona’s utterance with jo, these have apparently not led him to the same interpretation of the world as the one conveyed by Anne Mona’s utterance. Still, we will argue, this is not a counter example to the hypothesis that jo is a marker of mutual manifestness.

Even though the same set of facts are manifest to two individuals, there is no guarantee that this set of manifest assumptions will lead the two individuals to the same conclusions about the world (Blass Reference Blass, Andersen and Fretheim2000:45). Using a mutual manifestness marker is therefore only predicted to be infelicitous or marked if the addressee does not have the necessary evidence that would in principle enable him to accept the proposition as true or probably true. In the context of (11), Ramsey does have access to the necessary evidence: he and Anne Mona both live in the BigBrother house, and he thereby has perceptual access to her actions. He has, however, interpreted these as qualifying her for the labels sweet and kind. Anne Mona, on the other hand, does not perceive her actions as indicating that she is kind. Her use of jo is thus intended to help Ramsey realise this by reconsidering which memories about her actions to take into account, and to suggest a different judgment of these. He thus has the necessary evidence which makes him capable of drawing the – according to Anne Mona – correct conclusion with respect to her character. It can be rhetorically strategic to communicate that the addressee has access to information that enables him to accept a proposition p as true when the speaker knows that he holds the belief that ∼p. When Anne Mona uses jo in (11), she presents the proposition as mutually manifest, presumably because she wants Ramsey to realise that he actually is capable of accepting p as true. In this type of cases, where the addressee holds a different view than the speaker, jo may have a strong persuading effect because it encourages the addressee to rethink what evidence he has and to endorse p on the basis of that evidence.

The examples presented in this sub-section represent the various types of epistemic relationships between the hearer and the propositional content of the jo-utterance that we have found in the data set. These types of epistemic relationships are summarised in (12):

1. (12)

   

The situations described in (12a–c) are all compatible with Sperber & Wilson (Reference Sperber and Wilson1986/1995) definition of mutual manifestness. We do not mean to suggest that there are no other ways an assumption can be mutually manifest.Footnote ¹¹ However, the list in (12) may be helpful for the purpose of understanding in what contexts sentence-internal jo can be used, for discussions about what is meant by the notion ‘mutually manifest’, and as a starting point for cross-linguistic comparison of phenomena that may be sensitive to mutual manifestness.

Note that (12a–c) are roughly compatible with the alternative descriptions of jo’s givenness aspect in Fretheim (Reference Fretheim1991), Andvik (Reference Andvik1992) and Lind (Reference Lind1994). Andvik (Reference Andvik1992:50) acknowledges that jo is appropriate when the information is available through general cultural knowledge, through the discourse or extralinguistic context, and with information which the addressee might expect but not know to be the case. However, Andvik’s (Reference Andvik1992:86) label ‘consensus about the facts’ is slightly misleading when faced with cases like (11), where the addressee has explicitly expressed the opposite of the content of the jo-utterance. As for Fretheim’s (Reference Fretheim1991) term ‘mutually known’, this generalisation has to be supplemented with a list of exceptions to accommodate cases where the proposition is not known to the addressee, but has to be inferred or activated at the moment or during the processing of the utterance. Lind’s (Reference Lind1994) term ‘interpersonal’, in the sense of creating a shared context, seems compatible with the set of interpretations we have observed for utterances with jo. However, the label mutually manifest is preferred in our view, because it is a relatively strictly defined notion which applies to cognitive aspects of communication in general. It therefore brings more accuracy to a semantic proposal intended to reflect precisely how the meaning of jo is stored in the mental lexicon.

4.3 A semantic constraint on the proposition as a premise

Throughout the data set, the interpretations of jo-segments lead to the derivation of implicatures in addition to presenting the proposition as mutually manifest. One of the clearest examples is (2) in Section 2.1 above: Rodney’s use of jo in the utterance where he claims that he has seen Anette naked several times, not only presents this assumption as mutually manifest, it also signals that Anette should draw further conclusions from this assumption (see also Section 2.2 above). His utterance thereby arguably communicates the implicature that Rodney should be allowed to stay in the bathroom while Anette showers since he has already seen her naked several times. The line of derivation was explicated in (4) where the first premise is the propositional content of the jo-utterance, the second premise is a contextual assumption, and the conclusion is the implicature of the utterance. If we remove jo from Rodney’s utterance, the implicature would not be as easily accessible as it is when jo is used. Jo thereby decreases the processing effort necessary for obtaining cognitive effects, because it makes the route to the intended implicature more accessible.

Also (11) presented in the previous sub-section illustrates that jo encourages the addressee to look for implicatures: Anne Mona’s utterance receives an optimally relevant interpretation in the given context only if it is taken as a contribution to the discussion of whether or not she is likely to win the competition. Ramsey has just conveyed that she is likely to win, because the sweet and kind little girl typically wins. In this context, Anne Mona’s utterance makes the following line of inference highly accessible: it is mutually manifest that p ‘Anne Mona is not kind at all’, and therefore it is mutually manifest that she is not likely to win the competition. Thus, the segment with jo yields cognitive effects as a premise which supports the assumption that Anne Mona will not win the competition.

Also in cases like (6) and (8)–(10), it should be fairly clear that the jo-segment constitutes a premise. In (6), the presence of jo in ‘you have JO crumbs all over your sweater’ makes the following line of inference highly accessible: ‘Per Morten has crumbs all over his sweater; if one has crumbs all over one’s sweater one does not know how to eat properly; thus, Per Morten does not know how to eat properly’. In (8), the presence of jo suggests that the utterance is meant as an argument for why the speaker looked so much forward to moving into her new house. A possible line of inference is the following: ‘Since the speaker is born in Ravnkloa, moving to a house by the canal is almost like coming home; coming home is good; therefore, moving to a house by the canal is good’. Without jo, it would have been less clear that the utterance is supposed to provide a premise for this conclusion. In (9), the jo-segment is a premise that supports a contextually available assumption that electronic darts is a great invention, and in (10) the jo-segment supports the previously explicitly communicated assumption that the editor does not catch Vinje’s actual message.

In (7), the main function of the utterance with jo seems to be to remind the hearer about the meaning of demolition man. In cases like this, it may be less obvious how the jo-segment constitutes a premise. At a closer view, however, this property is in fact part of the interpretation: the content of the jo-utterance indirectly supports the assumption that the movie is a hilarious and hysterical action movie, since the title Demolition Man suggests that there will be a lot of random demolition. The line of inference is outlined in (13):

1. (13)

   

While the jo-utterances in our data set usually backwards support an already mentioned or available assumption (e.g. as in (6) and (8)–(10)) or support a new conclusion (e.g. as in (2) and (11)), (7) shows that a syntactic structure where jo occurs in an appositive relative clause can yield an interpretation where the content of the jo-clause indirectly supports the content of the main clause.

The implicated conclusion derived on the basis of the jo-segment is a particularly salient part of the utterance interpretation in contexts like (2) and (11). The utterances are relevant by virtue of communicating and supporting the implicated conclusion that the boys can stay in the bathroom and that Anne Mona is not likely to win the competition, respectively. Also in (6) and (8)–(10), the implicatures are salient parts of the ostensively communicated content. The analyses of cases like (7) show that the implicated conclusion based on a segment with jo can also be communicated fairly weakly, but nevertheless be an important part of the overall utterance interpretation: in order to get a maximum number of cognitive effects, the addressee of (7) needs to derive the assumption rendered in (13d). The effect of using jo in such cases is to signal that the content is to be used as background information that supports the communicator’s statement in the main clause. Throughout the whole data set, the very presence of jo appears to ease the derivation of the implicatures by means of signalling that the basic explicature is intended as a premise for deriving and supporting a contextually available conclusion, that is, an implicature. Sometimes the implicature is weakly communicated and figures in the background as in (7), but mostly it is strongly communicated as in the other data discussed in this sub-section.

4.4 A semantic constraint on opposition?

Andvik’s (Reference Andvik1992:61) claim that a jo-proposition is interpreted as a premise for deriving a conclusion is perfectly consistent with the data we have examined. However, the hypothesis that jo constrains the interpretation such that the conclusion always counters an assumption – or an ‘idea in the air’ in Andvik’s terms – is less clear. It is evident from our sample of utterances with jo that the implicature triggered by jo often yields cognitive effects by contradicting an assumption which is either entertained by the interlocutors (data like (6) and (11)), attributed to the interlocutors (data like (2)) or may be attributed to other individuals (data like (5), (8)–(10)). This is however not always the case.

In datum (7) discussed in the previous sub-section, the interpretation hardly includes an element of opposition, and there are no obvious contrasting views to the conclusion rendered in (13d). One may of course argue that a jo-utterance opposes potential objections. But intuitively, the purpose of explaining the meaning of the title Demolition Man is not to argue against anyone’s view – it is rather to provide background information which is necessary to interpret the utterance, and this includes understanding why this movie is a typical contemporary action movie. It therefore seems counterintuitive to claim that the speaker assumes that someone holds the view that Demolition Man is not a typical contemporary action movie. (14) is another example of this type:

1. (14) Hvor kommer vinden fra? Hvorfor blåser det, og hvordan skapes vinden? Professor Helge Nørstrud ved Institutt for mekanikk, termo- og fluiddynamikk vet svaret: - Det er solas oppvarming på jorda som skaper vind. Sola belyser jo jorda på forskjellig måte, og sammen med jordas rotasjon skapes det energi- og trykkforskjeller i lufta.

2. (Oslo Corpus)

3. ‘Where does the wind come from? Why does it blow, and how is wind created? Professor Helge Nørstrud at the Department of Mechanics, Thermo- and Fluid Dynamics knows the answer: it is the sun’s heating of the earth that creates wind. The sun shines on the earth in different ways, you know, and together with the earth’s rotation, this creates energy- and pressure differences.’

‘The sun shines on the earth in different ways, you know’

The reason for reminding the hearer of the information in the jo-segment is that it is among the premises in the explanation of how wind is created. According to our intuitions, the use of jo in this example does not suggest any more than a corresponding utterance without jo that somebody may have other beliefs about how wind is created. In other words, there is hardly an idea in the air which is countered by the explanation provided by the professor. Consider also (15):

1. (15) Etter en stund kommer komiteen opp med et antall kandidater. Dette er utelukkende guttebarn, ca. i fire-års alderen. Kandidatene blir deretter utsatt for en hel del prøver for at munkene skal finne “den rette”. Den rette - som jo er en reinkarnasjon av den avdøde Dalai Lama, vil kjenne igjen gjenstander når de vises noe som har tilhørt Dalai Lama …

2. (Oslo Corpus)

3. ‘After a while the committee arrives at a set of candidates. They are all boys around the age of four. The candidates are then exposed to a range of tests for the monks to find “the right one”. The right one – who is, remember, a reincarnation of the deceased Dalai Lama, will recognise artefacts that used to belong to the deceased Dalai Lama when exposed to these …’

‘who is, remember, a reincarnation of the deceased Dalai Lama’

In (15), the writer reminds the reader that ‘the right one’ is a reincarnation of Dalai Lama. This information is necessary in order to understand why the committee exposes the candidates to artefacts which belonged to the deceased Dalai Lama. More precisely, the jo-segment contributes a premise for deriving the implicature ‘the right one will recognise the artefacts that belonged to the deceased Dalai Lama because he is a reincarnation of the deceased Dalai Lama’. There is no reason to assume that an optimally relevant interpretation of the jo-segment requires the reader to entertain the assumption that someone might oppose to this causal relation. Note that the jo-facilitated implicature does not argue for the reality of reincarnation: rather, it serves as background information in the overall description of the committee’s work and strategies. A similar use of jo is noted by Solberg (Reference Solberg1990:67), who states that jo can mark background information that clarifies the connection between the surrounding segments. Oppositional aspects are not part of her analysis of this use of jo.

It should be clear by now that jo-facilitated implicatures do not always counter an idea in the air. There is one more example we would like to show, as the interlocutor’s utterances in this example provide explicit evidence that there is no available assumption to which the jo-implicature is in opposition. In (16), A and B are talking about how A never manages to open bottles of fizzy drinks, and that she is not that fond of drinking fizzy drinks either:

1. (16)

   

(NoTa–Oslo)

‘As we know, you always stir the fizz away.’

B’s utterance with jo communicates the basic explicature ‘you always stir the fizz away’, which contributes to relevance by supplying an argument that supports the conclusion ‘A is not fond of fizzy drinks’. Judging from the preceding discourse, this assumption is available and entertained as true by A and B alike, and the jo-implicature thus yields cognitive effects by confirming an existing view. Note that Andvik (Reference Andvik1992:79–80) and Lind (Reference Lind1994:108) would probably argue that there is indeed opposition in cases like (16). On their view, the conclusion ‘A is not fond of fizzy drinks’ would counter an assumed or expected opposition. On our Relevance-theoretic account, on the other hand, an attempt to identify an assumption that the jo-implicature could oppose in cases like (16) would obscure the analysis of the comprehension process and the speaker’s communication intentions: an assumption about a possible opposition is not worth accessing in the comprehension process. In other words, it would cost cognitive effort to entertain the assumption that someone might believe that A is fond of fizzy drinks. It is far less costly to go directly to the interpretation where the conclusion yields cognitive effects by confirming A and B’s shared assumption that A does not like fizzy drinks.

If the semantic proposal includes an opposition aspect, it would make false predictions about the interpretation of jo in (7), (14)–(16) and similar cases in our data set (14.37% of our data). On the other hand, in the majority of our data set (85.63%), it is either clear that the jo-implicature opposes a contextually available assumption, or it is possible to argue that the jo-implicature opposes an assumption which is attributed to the interlocutor or to an imagined third party, e.g. in terms of assumed mainstream beliefs about electronic darts or neighbourhoods in Trondheim (recall (8)–(9)). The semantic proposal and the pragmatic account of jo need to reflect that such opposition interpretations are frequent, without making false predictions for interpretations like those in (7) and (14)–(16). A solution with two lexical entries, that is, one for oppositional jo and one for non-oppositional jo, would be problematic. Such an analysis would predict that hearers have to choose between an oppositional and a non-oppositional variant when they interpret an utterance containing the form jo, but there seems to be no clear borderline. In cases like (8), for instance, there is arguably some opposition between the implicature and mainstream assumptions about neighbourhoods in Trondheim. On the other hand, the utterance may just as well be interpreted mainly as an explanation of why the speaker is thrilled to live close to the canal without the assumption that this explanation is intended as a justification for his preference.Footnote ¹²

5.1 The semantics and pragmatics of jo

Traditional Relevance-theoretic analyses of pragmatic particles and discourse markers have typically involved a single procedure, not several. These tendencies among Relevance-theoretic works are, however, not due to fundamental properties of Relevance Theory. As Wilson (Reference Wilson, Escandell-Vidal, Leonetti and Ahern2011) argues, there is no reason to assume that an expression could not encode both conceptual and procedural meaning. Given this, we see no reason why it could not also be possible for an expression to encode a procedure which consists of more than one procedural constraint (see Blass Reference Blass, Andersen and Fretheim2000, Reference Blass2012; Borthen Reference Borthen2014). The question of how many procedural constraints are involved should be an empirical question, and in the case of jo, the empirical investigation in Section 4 shows that two constraints apply. We therefore conclude that Norwegian sentence-internal jo has the following semantics:

1. (17) A procedural semantics for sentence-internal jo

   In an utterance that contains sentence-internal jo and is used to express the proposition p,

In accordance with the discussion in Section 4.4, we do not assume that jo encodes a third procedural constraint such that the implicature has to stand in opposition to another contextually available assumption attributed to the hearer or another person. The frequency of opposition interpretations nevertheless must be accounted for, and the explanation seems to be found in the type of meaning encoded by jo. More precisely, it lies in the type of procedural meaning activated by jo, which is, as we shall argue, linked to the sub-domain of the human cognitive system which Wilson (Reference Wilson, Escandell-Vidal, Leonetti and Ahern2011) labels the argumentation module.

The argumentation module is one of many modules that make up the human mind. Wilson describes the massive modularity hypothesis as follows:

The cognitive mechanisms in the respective modules are associated with capacities for e.g. parsing and speech production, mind reading, emotion reading, social cognition and argumentation. Wilson (Reference Wilson, Escandell-Vidal, Leonetti and Ahern2011) suggests that clusters of procedural linguistic items are linked to certain capacities, and their function is to put the hearer in a state where ‘some of these domain-specific cognitive procedures are highly activated (and hence more likely to be selected by a hearer using the Relevance-theoretic comprehension heuristic)’ (Wilson Reference Wilson, Escandell-Vidal, Leonetti and Ahern2011:6). Expressions like Wow! and Alas!, for instance, may be linked to procedures for emotion reading, and honorific expressions may be linked to the capacity for social cognition (ibid.:19–20). As such, they are linguistic cues that trigger the activation of certain modules.Footnote ¹³

Jo’s procedural meaning seems to be linked to the argumentation module, which is a set of cognitive mechanisms for epistemic vigilance. We employ this set of mechanisms to ensure that we are not accidentally or intentionally misinformed by others when we engage in communication (Sperber et al. Reference Sperber, Clément, Heintz, Mascaro, Mercier, Origgi and Wilson2010). Procedures for epistemic vigilance are divided into procedures for assessing the reliability of the source of communicated information and procedures for assessing the reliability of the content in terms of consistency and coherence. Wilson (Reference Wilson, Escandell-Vidal, Leonetti and Ahern2011, Reference Wilson, Walaszewska and Piskorska2012) proposes that discourse connectives (such as after all and so), evidentials, and epistemic modals are linked to mechanisms for epistemic vigilance because this group of expressions are used to mark the utterance content as part of an argument. More specifically, discourse connectives are linked to epistemic vigilance towards the content of the communicated information, in that they mark the logical relation between the proposition in their scope and other assumptions (Wilson Reference Wilson, Walaszewska and Piskorska2012). And epistemic expressions are linked to epistemic vigilance towards the source, in that they indicate the source and/or strength of the assumption represented by the proposition (ibid.; see also Berthelin Reference Berthelin2017:361–381). Arguments affect whether we believe something or not. When arguments are entertained publicly or privately, they affect our epistemic vigilance towards whatever assumption is up to evaluation. From this it follows that if a linguistic item marks the utterance as part of an argument, it plays a role in the epistemic evaluation of the assumptions at play and thus its use and comprehension is linked to the argumentation module.

With this in mind, let us consider the motivations a speaker may have for using jo (or a similar expression in another language) to indicate linguistically that the proposition is mutually manifest and should be used as a premise for deriving and supporting an available conclusion. In our corpus data, four types of contexts can be observed. In the list below, p represents the propositional content of the basic explicature in the jo-utterance, and q represents implicated conclusions:

1. (18)

   

All these motivations for using jo are connected to (i) the intention to strengthen the epistemic status of the proposition p and (ii) the intention to signal that p is a premise for deriving a conclusion. Thus, the use of jo is motivated by the intention to affect the hearer’s epistemic vigilance towards (i) the source of – and thereby the epistemic support for – the communicated information, and (ii) the content. Since p is mutually manifest, the hearer is encouraged to lower her epistemic vigilance towards the evidential source of the proposition – after all, she herself has access to the evidence supporting p. And since the implicature is a conclusion that follows from a mutually manifest premise, the hearer is encouraged to lower her epistemic vigilance towards the content of the utterance’s implicature. The use of jo thus appears to trigger the activation of the mechanisms for epistemic vigilance towards the source of the basic explicature and the content of the implicature. For this reason, jo is a useful tool when speakers suspect that the hearer will not accept the information they are communicating. This suspicion may spring from the assumption that the hearer holds the erroneous belief that ∼q, or that she will need some persuasion to derive and accept q, because she has either forgotten the premise p, or she holds the (false) belief that ∼p. The need for a linguistic device like jo is hence closely associated with the need to convince the hearer of the epistemic status of p and the epistemic status of implicatures q that follow from p (see also Blass Reference Blass, Andersen and Fretheim2000, Reference Blass2012).

The hypothesis put forward here that jo is linked to the argumentation module and triggers mechanisms for epistemic vigilance predicts the high frequency of contexts where jo is used to generate an implicature which opposes a contextually available assumption. At the same time, it does not follow that jo is only suitable when the intention is to persuade a particularly sceptical audience. Sometimes it can be useful to signal that the premise for an implicature is mutually manifest such that the hearer does not spend her cognitive energy on deciding whether or not to believe the premise and, in turn, fail to derive the implicature. This is the case in (15), for instance: The indication that the proposition p (‘the right one is a reincarnation of the deceased Dalai Lama’) is already manifest encourages the hearer to lower her epistemic vigilance towards p and the implicature q (the explanation of why they use these tests), and thereby use p and q as mere background information when she processes the overall utterance. And sometimes the lower epistemic vigilance is an epiphenomenon of the interpretations, as in (16) where the proposition p and the implicature q confirm the interlocutors’ shared assumptions.

5.2 Jo and mirative uses

The Norwegian major dictionary Bokmålsordboka (2017) states that sentence-internal jo can be used to highlight surprise (see also Solberg Reference Solberg1990:67). This meaning aspect is labelled mirativity in the linguistics literature and may be defined as the marking of ‘the utterance as conveying information which is new or unexpected to the speaker’ (DeLancey Reference DeLancey2001:369–370) as opposed to ‘knowledge which is already integrated into the speaker’s picture of the world’ (DeLancey Reference DeLancey2001:379). In our corpus survey, we found no examples with jo that conveyed a mirative meaning. However, consider the constructed minimal pair in (19):

1. (19) A and B are about to go skiing. A opens the door to go outside:

   

‘Oh look! There is no snow!’

A and B are about to go skiing, and hence they arguably hold the belief that there is snow. In this context, the b-sentence in (19) with jo is more natural than the a-sentence, since the b-sentence most clearly conveys the speaker’s surprise. As predicted by the semantics in (17), the presence of jo in (19) gives rise to an interpretation where p is mutually manifest and should be used as a premise for deriving an implicature. In (19), the implicature q (‘we cannot go skiing’) yields cognitive effects by contradicting the existing assumption (‘we can go skiing’). Interestingly, it appears that the speaker is also addressing himself in (19b), whereas (19a) merely informs the hearer about the propositional content. At a closer look, (19b) is similar to the type of data described in (18b) above. The only difference is that in (19), the speaker also used to believe ∼q (‘∼[we cannot go skiing]’), as he had not previously stored p (‘there is no snow’) in his mind. At the moment of the utterance, he accesses (visual) evidence that enables him to assume p, from which it follows that q.

We may now ask why a speaker would use linguistic means to refer to evidence which is obvious to the hearer in addition to herself in the context of the utterance. It follows from the communicative principle of relevance that in order for jo to be worth processing in such contexts, the speaker must intend to convey more than the mutually obvious fact that the evidence is mutually manifest by perception. When the reality of the evidence is obvious in the context of utterance, the speaker’s choice to mediate his statement by reference to evidence may well be due to his awareness of a discrepancy between his (and possibly also the hearer’s) current knowledge state, and the realisation of a new state or event (see Lazard Reference Lazard2001, Borthen & Knudsen Reference Borthen and Knudsen2014). We do not claim that any utterance with sentence-internal jo will convey mirativity if the evidence alluded to is physically present at the moment of the utterance (recall example (6) above). But if the speaker wants to communicate that she is surprised, the use of jo may yield the intended cognitive effects.