2.1 Token-reflexives

An occurrence of the proximal spatial indexical here is traditionally said to be token-reflexive if its reference includes the place where a given utterance with this word is being produced, whether the place is meant to be just the space occupied by the speaker, the neighbourhood where the speaker is at the moment of speaking, the province or the country in which the conversation is taking place, or the earth as opposed to other planets, etc. (see e.g. Braun Reference Braun and Zalta2007). We can imagine a token of (1) being produced as part of a telephone conversation.

1. (1) The weather is beautiful here.

The speaker of (1) is simply saying that the weather experienced at the moment is beautiful; the geographical boundaries of the region favourably affected by this weather will usually not be something for the addressee to worry about. Sometimes the referent of here may simply be identified as ‘where the speaker is at the time of utterance’, which we may call the context-insensitive, description-based denotation of any occurrence of here used token-reflexively. Interpreting the reference of here in an utterance of I can see two lakes from here does not involve any context-sensitive estimation of the speaker's referential intention, though it does require the hearer's perception of the location of the speaker. On the other hand, resolution of the reference of here in the question Is this your first visit here? depends on the hearer's consideration of how much space the speaker intends to include in the reference of the indexical – a hotel, a village, a country?

With the exception of the data to be discussed in Section 5, the token-reflexive here cannot be replaced by there, as the domain of discourse relevant to determination of the reference of there may not include the location of the communicating agent at the time of utterance.Footnote ² This distinguishes the token-reflexive use of here from its anaphoric as well as its deictic use, both of which allow the speaker to use there when it would also be acceptable to use here, but not vice versa. There is a special proximity requirement that any token of the English indexical here must fulfil, whether the requisite shortness of the distance is assessed in terms of perceived length or circumference, in terms of the textual distance between an anaphoric token of here and its inferred antecedent, or in terms of a more elusive mental distance within the communicator's ‘context space’ (Gundel, Hedberg & Zacharski Reference Gundel, Hedberg and Zacharski1993), as when the linguistically represented entity is a metarepresented thought presented as a part of a fictional text, and not a situated public utterance.Footnote ³

In (2) below, the intended extension of the term here used token-reflexively is constrained by the fact that the meaning of the verb causes us to immediately think of the speaker's dwelling as the most likely referent of here, though it is also conceivable that the speaker regrets the direction in which her country is heading politically, in which case here stands for something considerably more extensive than the house in which she lives, a plausible implicature being that she considers leaving the country for good.

1. (2) I simply can't live here any longer.

Speaker B's response to A's question in (3) also contains a token-reflexive here, but here it is natural to suppose that interlocutor A may have expected a more accurate description of the location of the cathedral.

1. (3) A: Where in this city is the cathedral?

2. B: The cathedral is here.

Speaker B's utterance does not mean that A and B are now inside the cathedral, though it is pragmatic considerations, not the linguistic semantics or logical form of B's sentence, which will normally exclude that interpretation. The token-reflexive use of here only requires that the space occupied by the interlocutors and the space occupied by the cathedral referred to are within the same circumference. What is not made clear by speaker B is whether she means ‘this neighbourhood’, ‘this side of the river’, etc., but speaker A's specification in this city constrains A's own search for the referent of B's indexical here.

In (3ʹ), speaker A's added disjunctive specification on this side or the other side of the river serves as a constraint on the referential interpretation of here in B's response. Thanks to A's afterthought specification of what the wh-word where is meant to signify, B can safely choose the proximal spatial term here with no risk of being misunderstood.

1. (3ʹ) A: Where in this city is the cathedral – on this side or the other side of the river?

2. B: The cathedral is here.

As we have no problem inferring that B's utterance in (3ʹ) means ‘The cathedral is on our side of the river’, the prepositional phrase on this side in A's utterance is seen to play the role of an antecedent providing information needed to saturate the reference of here in context. B's utterance in (3ʹ) exemplifies a token-reflexive use of here but at the same time it is also anaphoric, as its referential interpretation depends on an inferred link between the token of here and the earlier use of the proximal demonstrative determiner this of this side (of the river). However, it differs from the anaphoric use to be discussed in Section 2.2 below, in that its token reflexivity excludes the use of there as an alternative term. If B had said there instead of here in (3ʹ), interlocutor A would have a reason to believe that B intended to tell A that the cathedral is on the other side of the river. The jointly token-reflexive and anaphoric use of here in B's utterance in (3ʹ) also deviates from an anaphoric here which is not concomitantly token-reflexive, in that speaker B does not have to worry about what we are going to refer to as the speaker-activation condition on appropriate use of here, a discourse constraint that will be introduced in Section 2.2. Here is in order in (3ʹ), even though the antecedent phrase that is the most important determinant of its reference was not uttered by B but by interlocutor A.

We assume that speaker B in (3ʹ) intends the hearer A to handle the word here pragmatically both as a token-reflexive and as an anaphoric indexical, because this will lead A to successful resolution of its reference. Whereas distal there is the default English spatial indexical when it is the anaphor in an antecedent/anaphor relation, an anaphoric spatial indexical which is at the same time token-reflexive must obey the rule that token reflexivity demands the proximal indexical here. Following Powell's (2010) terminology, we will say that it is B's derivational intention in producing (3ʹ) that A undertakes a dual pragmatic processing of B's indexical and associates this token with the properties pertaining to spatial anaphora as well as the properties pertaining to spatial reflexives.

Apart from the fact that speaker B in (3ʹ) would have produced a less relevant utterance if she had made the extremely unlikely decision to topicalize the spatial indexical and produce the sentence Here is the cathedral instead of The cathedral is here, B did not exploit any linguistic device that would have overtly signalled to A how here in (3ʹ) should be treated in the pragmatic process of saturation of the indexical, that is, determination of its semantic value in the propositional content of the utterance (e.g. Recanati Reference Recanati2004). There is no specific procedural clue in B's utterance which is intended to inform A of B's derivational intention, no overt indication that here is to be processed as a token-reflexive and at the same time an anaphoric indexical. A will understand B's derivational intention on the basis of accessed contextual assumptions rather than procedural information encoded in B's verbal stimulus. In Section 3 it will be demonstrated how the inverted word order illustrated by a sentence like Here is the cathedral does provide procedural information about the speaker's intention to signal that resolution of the reference of here must rely on the inference that the indexical is used for a deictic purpose. An utterance-initial here is often less suitable for token-reflexive reference than a token of here in a post-verbal position, though contextual matters may occasionally neutralize or nullify the normal procedural effect of the difference between a pre-verbal and a post-verbal position for here (cf. examples (6) and (7) below, where here is fronted but token-reflexive).

In (4) the use of here is strictly token-reflexive but the geographical area referred to is restricted by the disjunct (or) west of here.

1. (4) Is the cathedral here or west of here?

Here and west of here offer only a rough indication of the intended extension of here, but the second term of the disjunction divides a diffusely delimited geographical area into two complementary parts, thus restricting the reference of here westwards from the spatiotemporal vantage point of the interlocutors.

There are exceptions to the rule that the reference of a token-reflexive here must coincide with the location of the communicator. The possible existence of a note posted on an office door which reads as shown in (5) – an example discussed by Recanati (Reference Recanati2010), among others – shows that token-reflexive here does not require the presence of the communicator at the place referred to at the time when the message is received and interpreted.

1. (5) I am not here now.

Now and here in (5) must be interpreted as indexicals whose reference is respectively the time of reception and the place of reception of the message, not the time and place at which the message was produced (which would make the utterance of (5) self-contradictory). The place referred to is the place where the post-it containing (5) is, but no communicative act is performed until some individual, preferably an intended addressee, reads the note. Fillmore (Reference Fillmore1997:83) makes the point that here cannot normally refer to where the addressee is unless the communicator is there too, but the written negative declarative of (5) posted outside an office is a counterexample to that generalization.

Here may be used to refer token-reflexively to that point in the text where the word occurs. It is then a meta-discursive pointer to a mutually manifest set of contextual assumptions at the time of utterance, something that might be paraphrased as ‘at this point in the current discourse’. The sentence-initial position is then quite normal; the indexical serves as a link between the context retrieved on the basis of the preceding discourse and the content of the sentence it introduces. In the case of (6), one might consider classifying the fronted token of here as a cataphoric indexical, as the utterance suggests that the addressee has the right to anticipate that the immediately following discourse will include a demonstration of the sort described.

1. (6) Here it will be demonstrated why we have to be extremely cautious.

Here in (6) would be an anaphor if its referential interpretation depended on access to some antecedent information, an assumption that the predicate will be demonstrated rules out, but is here a true cataphor in (6)? (7) illustrates a use of here that we intuitively equate with its use in (6) but there is no ‘postcedent’ specification of what here is supposed to entail in (7), so here cannot be cataphoric in (7).

1. (7) Here one might have expected to find a demonstration of why extreme caution is necessary, but no such demonstration is offered.

We believe there is no need for a special ‘cataphoric’ subcategory of uses of here. The indexical is token-reflexive in (6) and (7) alike. That here may be linked to subsequent information in a discourse that includes an utterance of (6) is a consequence of the future temporal reference indicated by the expression will be demonstrated.

The proximal spatial indexical in the West African language Ga is biε, which may be used sentence-initially with the meaning of ‘at this point in the discussion’. One attested example is (8) (where the lengthy content of the response is immaterial and is therefore left out).

1. (8)

Ga biε ‘here’ refers token-reflexively in (8) but there is also a signalled link to the immediately preceding part of the discussion, and we get the impression of a causal link between what the chief responds to and his response. This makes the use of biε in (8) rather similar to anaphoric uses of here and we are tempted to conclude that the proximal spatial indexical in this example is on the one hand token-reflexive but on the other hand anaphoric.

Due to the inextricability of a speaker's place and time coordinates, an expression that literally refers to the place here-and-now will sometimes be found to substitute for what would appear to be more logically represented by a temporal term, in written as well as spoken discourse. Metaphorical use of spatial terms for temporal reference appears to be a very widespread phenomenon, as the following two statements confirm (see also Haspelmath Reference Haspelmath1997):

Here in the utterance of (9), produced by a teacher who finds it necessary to end her class on this note, is token-reflexive, but the reference is temporal, not spatial, in spite of the lexical meaning of the indexical.

1. (9) Let's stop here.

When a spatial indexical is employed for temporal reference, it is invariably of the token-reflexive sort.

Hungarian discourse is characterized by an extensive token-reflexive use of spatial itt ‘here’ where other languages might prefer a temporal term because the reference point is clearly temporal. The following is a sentence appearing in the Hungarian translation of a Norwegian novel. The direction of the arrow (←) indicates that the Hungarian sentence is a translation of the Norwegian sentence.

1. (10)

The first person protagonist is relating what he is experiencing at a particular moment. Conceptually, the emphasis does not seem to be on the place where he is situated. Nevertheless the most idiomatic Hungarian rendering of Norwegian nå ‘now’ requires what appears to be a linguistic shift from the temporal domain in Norwegian to the spatial domain in the Hungarian translation. The same phenomenon is illustrated by (11), from the same novel as (10). The reader is invited to follow the man's temporally successive moves, step by step, and at the point where (11) occurs, he informs the reader that he has just entered the Volvo showroom, and yet the proximal spatial term itt, rather than its proximal temporal counterpart most ‘now’, is the preferred choice in Hungarian.

1. (11)

It was natural for the translator to change from the temporal indexical in the source text to a spatial one in the translation, in spite of the fact that the later phrase a Volvo-üzletben ‘in the Volvo showroom’ literally specifies the location of the protagonist. This seems to make the spatial term itt cataphoric but the derivational intention of the translator is that the reader should interpret the reference of itt as temporal in spite of the lexical meaning of this indexical. Hungarian usage makes this mental shift to a temporal reference unproblematic, while the corresponding Norwegian sentence structure of ?Her står jeg i Volvo-butikken ‘Here I am standing in the Volvo showroom’ would be severely unidiomatic.

As noted above, the use of a proximal spatial indexical like here for what should be interpreted as reference to the time of utterance presupposes token reflexivity. When here is used as an anaphoric term or as a deictic pointer to a sensory image, it cannot ever be replaced by now, and while here as anaphoric or deictic term may be replaced by there, the substitution of there for here is impossible not only in cases of regular spatial token reflexivity but also when the intended reference of the spatial indexical must be sought in the temporal rather than the spatial sphere.

2.2 Anaphoric use of ‘here’

Both proximal here and distal there can be used as anaphoric terms whose context-dependent semantic value depends on how the anaphor is saturated, that is, on the addressee's retrieval of antecedent information needed in the reference resolution process. There are no special grammatical constraints on the use of distal there when it is used as a pro-adverbial anaphor. When it is an anaphoric term, there does not contrast paradigmatically with here but is a distance-neutral indexical. There in (12) and the phrase in it in (13) can be used interchangeably as long as the first sentence includes a nominal constituent, here the second short story, which is seen to fulfil the role of discourse antecedent enabling the addressee to saturate there and it, respectively.

1. (12) The second short story is probably the best one. There's a kind of intensity there which is lacking in the others.

2. (13) The second short story is probably the best one. There's a kind of intensity in it which is lacking in the others.

Due to the topicality of the short story referred to and the temporal proximity of the antecedent and the spatial anaphor in (12), it is also possible to use proximal here anaphorically. The example in (14) includes an alternative linguistic form that seems to impart the communicator's personal engagement better than the unmarked anaphoric indexical there in (12), though the truth-conditional content is the same in (12), (13) and (14).

1. (14) The second short story is probably the best one. There's a kind of intensity here which is lacking in the others.

A prerequisite of the acceptability of proximal here in data like (14) is that here and its antecedent are produced by the same communicator, what is called the ‘speaker-activation condition’ on the use of proximal demonstratives in Gundel et al. (Reference Gundel, Hedberg and Zacharski1993). These authors were primarily concerned with nominal demonstratives like proximal this/these vs. distal that/those, but we will show in Section 4 that their speaker-activation condition applies equally to the proximal adverbial indexical here in many languages. The condition is not met in (15), which is why B₁ with there is in order but B₂ with here is felt to be deviant by some native speakers of English.

1. (15) A: The second short story is the best one, don't you think?

2. B₁: Yes. I agree. There's an intensity there which is lacking in the others.

3. B₂: Yes. I agree. (#)There's an intensity here which is lacking in the others.

Observe that the speaker-activation condition does not apply to the token-reflexive use of the proximal spatial indexical. This is why the use of here is in order in the utterance of (3ʹ) B in Section 2.1 above.

The speaker-activation condition on the use of proximal demonstrative anaphora seems to be more strictly observed in some languages than in some others. Russian (Gundel et al. Reference Gundel, Hedberg and Zacharski1993) is a language that is said to tolerate the use of proximal demonstratives to a larger extent than English when one speaker activates the referent and another speaker refers, anaphorically, to the same thing a moment later. If this is correct, we witness here a property of Russian that may shed light on a potential difference between the patterns of distribution of the spatial indexicals in the Slavic languages compared to the distributional ranges of the corresponding indexicals in most other European languages, an issue that will be discussed in Section 4.

The use of here in the following attested, written example appearing in Predelli (Reference Predelli1998) is anaphoric, but here – another example of the anaphoric use of the term here! – we are invited to accept the make-believe that the writer is actually reporting from a position in the Mount Washington district.

1. (16) If an entire neighbourhood could qualify as an outdoor museum, the Mount Washington district would probably charge admission. Here, just northwest of downtown, are several picture-book expressions of desert culture within a few blocks.

   (Predelli Reference Predelli1998:407)

The reader of (16) is caused to undertake a mental journey from the page in the text that includes (16) to what is described as the Mount Washington district. This is an example of what has sometimes been called ‘Deixis am Phantasma’, a concept introduced by Bühler (Reference Bühler, Robert and Klein1934).

In (17) is an announcement heard by London Underground's Victoria Line passengers just before the train stops at Warren Street Station.

1. (17) The next station is Warren Street. Change here for the Northern Line.

The voice could have said there instead of here, but proximal here is the standard form for all announcements of next stops in the London Underground system. If one is a passenger who needs to change from Victoria Line to the Northern Line at Warren Street, then the Warren Street station has a special significance for that person. The advice Change here . . . in (17), in contrast to the neutral anaphoric indexical there of Change there . . ., underscores the special salience of the referent Warren Street Station for those passengers who need to change to the Northern Line, a procedurally conveyed salience that dovetails with the descriptive meanings of the declarative and the following imperative in (17). The word here is no token-reflexive indexical there, because the intended referent of here is not the place where the speaker or the hearer is at the time when the voice producing (17) is being heard. Resolution of the reference of here depends on the hearer's interpretation of the meaning linguistically encoded in the preceding sentence, so here is an anaphor saturated by the previously given information that the next station is Warren Street. At the time of utterance, the train must be somewhere between Oxford Circus and Warren Street if the train is northbound and somewhere between Euston and Warren Street if it is southbound. Less than half a minute later, when the train comes to a standstill at Warren Street Station, the same voice announces: This is Warren Street Station. Change here for the Northern Line. This time the use of here is token-reflexive, like this in the preceding utterance, and may therefore not be replaced by there, but it is nevertheless also an anaphoric indexical whose antecedent is the noun phrase Warren Street Station in the preceding sentence. When the train is in motion and approaching Warren Street Station, as in (17), the reference is strictly anaphoric, not at the same time token-reflexive.

3.1 The relevance of the syntactic position of here

It will be shown that the position of here in a sentence typically offers procedural information that facilitates the hearer's inference about the communicator's derivational intention, as defined in Section 1.2 above. Specifically, a token of here (and corresponding spatial indexicals in many other languages) that appears initially in an utterance prepares the hearer for a resolution of its intended reference that depends on the interpretation of an extra-linguistic sensory stimulus. This is what is meant by ‘deictic’ reference. When the syntactic position of here is post-verbal, a contrastive stress on this indexical is needed in order to alert the hearer to the fact that a deictic interpretation is intended, in spite of the fact that a post-verbal position otherwise indicates token-reflexivity or anaphoricity.

The deictic use of here is not reducible to either the reflexive or the anaphoric use of this proximal spatial indexical.Footnote ⁴

According to Corazza (Reference Corazza2004:6), ‘when “now” and “here” are not used as indexicals picking out the time and location of the utterance, they work in an anaphoric way’, and according to Corazza (Reference Corazza2004:156), ‘they [i.e. the terms here and now – our insertion] do not have referential independence; they work as anaphors inheriting their value from an antecedent to which they are linked and coindexed’. We feel that this does not tally with Corazza's view, reported in Section 1.3 above, that the interpretation of demonstratives is perception-based. A deictically used spatial indexical is perception-based, an anaphoric one is not. It would surely be possible to define ‘anaphoric’ in such a way that any occurrence of here which is not of the reflexive sort is anaphoric in the sense that its referential interpretation depends either on information found in a linguistic antecedent (endophoric reference) or on information based on extra-linguistic perception (exophoric reference). Only the latter type can be signalled by means of a deictic gesture in conjunction with the indexical item, and sometimes the indexical may not be interpretable unless there is an accompanying gesture. A pointing gesture in the direction of the spatial referent conveys what derivational intention governs the speaker's use of here. It offers the hearer the procedural information that the referent must be identified through perceptual recognition of the target of the pointing. Resolution of the reference of an anaphoric here that depends on recognition of a discourse antecedent is a different cognitive process than resolution of the reference of a deictically used here that rests on interpretation of an extra-linguistic stimulus. A terminological conflation is therefore not advisable.

Inverted word order with a fronted here in a copula sentence is a procedural device that may be used in order to direct the hearer to a deictic as opposed to a reflexive interpretation. While the utterance-final position of here in (18) below supports a reflexive interpretation, the utterance-initial here in (19) will normally serve as a deictic pointer to where the spatial referent is. There is a potentially important information-structural difference between the word orders of (18) and (19), which may even affect the truth-conditional contents of utterances of these sentence forms.

1. (18) The cathedral is here.

2. (19) Here's the cathedral.Footnote ⁵

We are deliberately using the cautious expressions ‘potentially important’ and ‘may . . . affect’, because the speaker's choice of intonation is bound to be a key provider of procedural information about how the spatial indexical is supposed to be processed. A contextually appropriate prosodic signal may even override whatever procedural information could otherwise be derived from the syntactic configuration of elements. It is important that the verb is the copula, which freely permits the subject–verb inversion pattern shown in (19), which again causes the subject noun phrase to end up in the sentence-final focus position.

The pragmatic processing required for the right interpretation of a token of (19) is very different from what is needed to understand the thought conveyed by a token of (18). The post-verbal subject phrase the cathedral is not an utterance topic in (19). There is no topic in (19), because an utterance-initial here in a predication with the English verb (to) be will invariably be interpreted as an all-news information structure, as a thetic judgment as opposed to the categorical judgment of (18) (Kuroda Reference Kuroda1972Reference Kuroda1992; Lambrecht Reference Lambrecht1994).

Compare the initial position of here in (20) with its final position in (21).

1. (20) Here are the members of the board, #but they are in the west wing of the building.

2. (21) The members of the board are here, but they are in the west wing of the building.

The clausal conjunct after but is odd in (20) because the implicit information we receive there is that the people referred to are out of sight at the time of utterance, and this is inconsistent with the procedural information conveyed by the fronting of here in combination with the inverted word order in the first conjunct. Here are the members of the board displays a constituent order that instructs the hearer to treat here as a deictic indexical that requires the visual presence of the people described as members of the board. The same ‘but’-conjunct is fine in (21), though, because the final position of here in (21) directs the hearer's attention to the default token-reflexive interpretation of the indexical, which implies that the referent of here is the entire building, not just the room where the conversation is taking place. In (21) the extension of here encompasses the west wing as well as the room where the interlocutors are at the time of utterance.

An utterance of Here are the members of the board, but they are in the west wing of the building could conceivably be a relevant stimulus if the speaker is pointing to a group of people in a photo, where the background details are such that the room might easily be mistakenly believed to be in a different part of the building than the west wing. The utterance would then be a correction of the addressee's belief about where the photo was taken: the location is the west wing, despite appearances. Here is in the initial position, which indicates that the indexical is not token-reflexive but a deictic term that refers to what can be seen in the photo.

Having now shown that the speaker's placing here at the beginning of the copula sentence gives the hearer the procedural information that the word is to be processed pragmatically as a deictic indexical, we see that it would be rash to conclude that deictic reference with here is invariably signalled by an accompanying gesture that is intended to delimit a specific visible space as the referent of the term. The pointing index finger is meant to facilitate the hearer's pragmatic processing of an occurrence of here and computation of its reference, but the co-occurrence of the indexical with a particular syntactic form, like ‘Here is X’ (distinct from ‘X is here’), can reveal the speaker's derivational intention without the extra help of a gesture. What matters more than anything else is that the speaker whose derivational intention is to guide the hearer to a deictic processing of the indexical is looking in the direction of the place referred to, and a bodily gesture is often a convenient way to buttress the thought that information about the space which includes the referent of here is where the speaker is looking, and pointing.

The kind of information conveyed by means of (segmental or non-segmental) formal devices of the strictly procedural sort is information that is typically not mandatory to avoid violation of a rule of grammar, but when it is left out there is a risk that the hearer will misunderstand the speaker's intention or that the pragmatic processing of the linguistic stimulus and computation of its intended cognitive effects will be more effortful. Thus, if the speaker intends the word here to be processed deictically and the word is nevertheless put in the sentence-final position that disfavours a deictic interpretation, the lack of relevant procedural information in the syntactic form must normally be counterbalanced by the presence of a demonstrative gesture that coincides in time with the production of the spatial indexical. No matter whether there is, or is not, a deictic gesture that accompanies the production of the indexical, identification of the referent always depends on recognition of an extra-linguistic stimulus and the speaker's ostensive direction of the hearer's attention to this stimulus.

The difference in procedural information between a sentence-initial yet non-topic token of the proximal spatial indexical and a sentence-final token may lead the hearer to major truth-conditional differences that even have an impact on the hearer's comprehension of the truth-conditional content of the rest of the sentence. Consider the syntactic differences between the Norwegian sentences in (22) and (23a–b).

1. (22)

1. (23)

In (22), we have an example of deictic use of the indexical. An utterance of (22) would be appropriate at a moment when Jens, a physical object, has caught the speaker's eye (the speaker and the hearer may have been waiting for him to arrive), or when the speaker is pointing to Jens in a picture. Her ‘here’ in (23a) will be interpreted as a token-reflexive indexical. The hearer will interpret the verb ha ‘have’ as ‘keep’ if the context of utterance permits this reading, or possibly even ‘detain’. The direct object Jens is topicalized in (23b). Depending on contextual input the sentence-final her in (23b) will be processed either as a token-reflexive or a deictic term, but the sentence-final position of her in (23b) does not strengthen the inference that the speaker's derivational intention is for the indexical to be processed as a deictic term. The sentence-initial position of her in (22) is an important procedural indicator of the speaker's derivational and informative intention when the physical appearance of Jens is new information not only to the hearer but also to the speaker who is describing what she has just observed. In contrast to how we interpret (22), the syntactic topicalization of Jens in (23b) makes it unlikely that the appearance of Jens has just now impinged on the speaker's perceptual apparatus and has caused the speaker to describe what she is experiencing. All of (22), (23a) and (23b) will be relevant stimuli in distinct sets of contexts.

Locative predications that involve the use of the contrasting distal demonstrative there behave in much the same way as proximal here in terms of the procedural information provided by the syntactic position of the indexical. As long as there is sufficient linguistic and gestural evidence that the speaker has the derivational intention that there be processed as a deictic indexical, the sensory accessibility condition on deictically used spatial indexicals, which requires that the speaker, but not necessarily the hearer, perceives the place referred to, will even hold for the reference of the utterance-initial distal there. The predication in (24) is about a uniquely identifiable referent described as a cathedral. The referent of there in (24) is a location some distance from where the interlocutors are, and this location is not necessarily identifiable through direct perception. The referent of the utterance-initial there in (25), however, is a perceived object, or possibly a pictorial representation of it.

1. (24) The cathedral is there.

2. (25) There's the cathedral.

An utterance of (25) is an all-news utterance which normally demands that the cathedral is visually present, even if it may be a kilometre or more away from the place of utterance. Alternatively the cathedral may only have a map-scale presence, its location being represented as a spot on a city map where the map designer has placed a church icon. The reference of there in (24) is deictic if the space referred to is visible to the interlocutors and the indexical is supported by a gestural demonstration, but if there in the utterance of (24) is de-accented, after an accented subject noun phrase, then it is most likely to be an anaphoric item, the reference being to a discourse-given place to be identified via information found in an antecedent phrase. Obviously, use of deictic there as opposed to here is disallowed when the speaker's derivational intention is for the hearer to realize that token-reflexivity combines with deixis.

The declarative after the indicated pause in (26) is an attested example uttered by President Barack Obama at a press conference on 10 December 2009, the day he received the Nobel Peace Prize in Oslo (the Norwegian television channel NRK1). It illustrates the procedural effect of a token of here positioned before the subject and the finite verb in a copula sentence. The straight word order of the alternative sentence form of You are here, Ben would have been truly odd.

1. (26) Is Ben here?

   (then five seconds of visual search across the room for the American journalist that President Obama has just referred to as ‘Ben’)

   Oh, here you are, Ben!

Notice the referential difference between the first and the second token of here in (26). The token of here in the interrogative Is Ben here? refers reflexively to the room where the press conference is taking place. Exactly two questions were permitted to be given to the Nobel Laureate, one by a Norwegian and one by an American journalist. When the Norwegian had asked her question and received the President's answer, Mr. Obama was apparently trying to spot one specific person among the audience, the second person who had been selected in advance to be given the opportunity to address the Nobel Laureate. The President was evidently familiar with the American journalist who had been given this honour and with the fact that he had been selected. The sentence-initial here in the utterance following the pause, however, offers the procedural information that it is not token-reflexive but deictic.

Here in the utterance Oh, here you are, Ben! could have been replaced by distal there: Oh, there you are, Ben!. As the TV camera was facing Mr. Obama at the rostrum, the utterance was accompanied by a simultaneous indexical pointing in the direction of someone the viewers of the television programme could not see, not until one camera was turned towards this man as he started to speak. The addressee of the utterance Oh, here you are, Ben! is this American journalist, and the speaker's use of the proximal term here as a deictic indexical is motivated by the fact that he had now been spotted.

The difference between the linguistic expression of spatial token-reflexivity and the expression of spatial deixis is not limited to the syntagmatic difference between a fronted here and here in a post-verbal position. In (27) we have a straightforward example of here used as a token-reflexive term. The basement referred to is where the conversation is taking place; (28) on the other hand has a syntactic form that supports a deictic processing of the indexical.

1. (27) We used to have scout meetings here in the basement.

2. (28) We used to have scout meetings in the basement here.

The prepositional phrase in the basement constrains the set of possible interpretations of the reference of the preceding token of here in (27). Grouping the indexical and the descriptive phrase in the basement together as shown in (27) gives the hearer the procedural information that in the basement is a modifier of here and that the speaker's derivational intention is that here be processed as a token-reflexive term: the interlocutors are in the basement at the time of utterance. Here in (28), however, is a modifier of the descriptive phrase in the basement with which it forms a syntactic constituent. An anonymous reviewer reminded us that, if we front the adverbial phrase in the basement in (28), here must be fronted together with it, which proves that in the basement here is a single constituent. Here in (28) specifies which perceptually accessible basement the speaker has in mind. The adverbial adjunct conveys the procedural meaning that, at the time of utterance the interlocutors are not located in the basement referred to. The interlocutors are likely to be outside the house whose basement the speaker is directing the hearer's vision to, presumably with a concomitant pointing gesture that reveals the speaker's derivational intention to make the hearer process here as a deictic indexical that delimits the visible space singled out for reference sufficiently for successful identification of the referent.

The syntactic difference between (29) and (30) does not appear to be directly comparable to the difference between (27) and (28). While the grammatical and semantic relationship between the cathedral and here in (30) seems to be of the same kind as the relationship between the basement and here in (28), the sentence-final here in (29), which is not grouped together with the cathedral, does not correspond syntactically to anything found in (27) or (28).

1. (29) The cathedral is quite something here.

2. (30) The cathedral here is quite something.

A cathedral is something one will only find in a city, whether or not it has a sizeable population. In (29), where the position of here is sentence-final, the token-reflexivity interpretation seems to be the only one accessible. We understand the extension of here to be the city where the interlocutors are located. In contrast to this, the role of here in (30) may be to constrain the hearer's search for the referent described as a cathedral to a site that is close by, maybe a couple of blocks away, in the very neighbourhood where the interlocutors are now walking. In contrast to its procedural meaning in (29) where the sentence-final position of here and its independence as a syntactic constituent virtually secures a token-reflexive interpretation, the syntactic position of here in (30) makes it a post-nominal modifier just as in (28). Depending on the actual physical setting, the post-nominal position of here in (30) may cause the hearer's perceptual attention to be directed to a visible building, typically in conjunction with a gesture.

Due to our knowledge of the world, the noun cathedral in (30) does not have the same impact on the reference resolution process as the noun basement in (28), so the hearer's decision to identify here in (30) either as a token-reflexive or as a deictic indexical is a direct consequence of the hearer's pragmatic calculation of what contextual premises yield the most relevant interpretation of the utterance. What is indisputable is that the syntactic positions of here in (29) and (30) do not give the hearer the same procedural information about the speaker's derivational intention. While an utterance of (29) supports a token-reflexive interpretation of here and rules out a deictic interpretation, an utterance of (30) supports a deictic interpretation, but if there is no simultaneous pointing, a token-reflexive result may be computed instead, provided there is no contrary contextual evidence.

On our understanding of the concept of deixis in relation to the proximal indexical here, this word is a deictic term in (30) if it is accompanied by a pointing gesture in the direction of the site of the cathedral, but a pointing gesture is not a necessary accompaniment of absolutely all instances of the deictic use of here. A necessary condition is that it is possible to perceive the object referred to from the place where the utterance is produced. In (29), in contrast to (30), the hearer's inference about the extension of here will not lead him to anything smaller than the city as referent. In particular, the reference is not constrained to some visible object as figure in a figure–ground relation, and it is therefore not of the deictic sort. It is difficult, in fact, to imagine a context that makes a manual gesture a natural accompaniment of the token of here in an utterance of (29).

There is a conspicuous formal difference between how speakers of Hungarian may indicate that the proximal spatial term itt ‘here’ is to be interpreted either as a token-reflexive or as a deictic indexical. If a Hungarian wants to express that the cathedral of the local diocese is a stunning piece of architecture, one suitable linguistic form with itt is the sentence in (31).

1. (31)

However, if the Hungarian speaker's intention is to convey the same as when the English definite description the cathedral here picks out a perceived building, she may add the adjective-forming suffix -eni to the stem itt-, resulting in the prenominal adjectival form itteni placed between the definite article az and the noun katedrális. This particular adjective provides the hearer with the procedural information that the location referred to is an attribute of the object described as a cathedral.

1. (32)

The word itt, without the suffix -eni, covers token-reflexive, anaphoric and deictic uses of the proximal spatial indexical in Hungarian, but the formal differentiation between itt in (31) and itteni in (32) may be said to reflect the cognitive significance of the distinction between spatial reflexivity in the former and spatial deixis in the latter.Footnote ⁶

3.2 The dual function of here as a token-reflexive and a deictic indexical

In a sentence like (28) above, the basement referred to is visible from the place of utterance, and the sight of the object picked out is what matters in most instances of deictic reference with here. However, senses other than vision may be important, or even decisive. The perceived source of information may be auditory. Returning to English (30) or Hungarian (32), we can imagine that the chiming bells of the tower of the cathedral tell the hearer approximately where the cathedral is. The edifice may not be visible from the place of utterance.

The hearer's identification of the referent of here in (33) depends on non-linguistic (perceptual) information derivable from an auditory stimulus. The utterance in (33) may have been produced in response to the addressee's previous utterance of Where are you? or I can't see you.

1. (33) I am HERE!

The use of here in (33) is token-reflexive, because it does refer to the place where the speaker is. Why, then, is (33) not tautological, given the fact that the proposition expressed appears to be necessarily true? (33) is informative because the indexical is not just token-reflexive; this is at the same time an example of the deictic use of here, as the hearer's auditory impression of where the voice is coming from is the extra-linguistic demonstration necessary for correct reference resolution. Even if the hearer can see the speaker of (33) simply by turning around, the speaker's voice sample is the relevant perceptual clue that causes him to turn around. The sheer acoustics of the auditory stimulus provides procedural information without which the utterance would not be relevant to the hearer. Indeed, the speaker could have achieved the same by yelling Hullo! or using the addressee's name, or even emitting a non-linguistic exclamation instead of producing the full sentence in (33). When here refers to a narrowly delimited space identifiable via auditory perception, the voice itself substitutes for the gestural demonstration typically needed to identify the referent of here when the source of information is a visual stimulus.

Taking stock, we may say that the sound of a voice serves a function similar to that of gestural pointing in an act of deictic reference. We do not agree with Kaplan's claim (Kaplan Reference Kaplan, Almog, Wettstein and Perry1989:491) that here is a ‘pure’ indexical in an utterance of I am in here cited at the beginning of this paper. Our position is that here is deictic as well as token-reflexive when the interlocutors cannot see one another, because one would only say I am in here when it is the acoustic signal that reveals where the speaker is. This is an exact parallel to the church bells example and the illustration in (33). It involves the hearer's identification of a perception-based source of information similar to what happens when a pointing index finger in an act of deictic reference directs the hearer's attention to a visual stimulus. The acoustic image produced by the voice is the sensory input that enables the hearer to locate the speaker, which is the entire point of uttering I am in here.

The fact that the proximal indexical in (33) is deictic and at the same time a token-reflexive term explains why here in (33) cannot be replaced by there and why I am in here cannot be replaced by I am in there in an utterance addressed to someone in an adjacent room who cannot see the speaker. This is impossible because the speaker is surely referring reflexively to the place of utterance. In contrast to this, a token of here which is deictic but not token-reflexive may be replaced by distal there without changing the conditions for spatial referent identification, provided the referent is not too close to the speaker. One can leave a newspaper on a table for the hearer to pick up and read, saying either Here's the newspaper or There's the newspaper, because the referent is the spot now being occupied by the newspaper, in conjunction with a word order that supports a deictic interpretation.

We have emphasized that the sentence-final position of here fails to offer the hearer the procedural information that a deictic interpretation is intended, so why do we classify the use of the sentence-final here in (33) as not only token-reflexive but also deictic? Why is (33) the normal syntactic form when the hearer cannot see the speaker? Why is use of the inverted word order of Here I am unnatural in that context? Data like (33) shows that our generalization about the procedural information conveyed by the contrast between a pre-copular and a post-copular here is in need of modification. The generalization expressed earlier is valid only when the spatial perceptual target is visible to the hearer. It may be counterproductive to put here in the initial position when the perceptual impression important for reference resolution is based on hearing rather than vision. To what extent these general observations are valid for language codes cross-linguistically is not known to us, but similar constraints on the use of a sentence-initial proximal spatial indexical as a guide to pragmatic processing do exist in other Germanic languages than English.

The example in (34) is a caption accompanying a photo that appeared in a Norwegian newspaper article.

1. (34)

Although this is deictic use of Norwegian her ‘here’, there is no pointing, obviously, as the text is meant to direct the reader to the image and to a designated person represented in the image. The description uniquely singles out the right individual, Jean Cocteau, and the photo uniquely delimits the perceptible domain of reference, just as the visibly present map is the domain of reference when someone touches a map of Poland with her forefinger while producing an utterance of the sentence Krakow is HERE (categorical statement with a topic followed by a comment) or Here's KRAKOW (thetic statement with no topic).

One striking property of photo captions like (34) is that proximal here may not be replaced by distal there, even though the proximal indexical is used deictically. We observed that Barack Obama could have said Oh, there you are, Ben! instead of Oh, here you are, Ben! when spotting Ben in example (26). Is it just proximity that guides the producer's choice of proximal here at the expense of distal there in the caption presented in (34)? No, what distinguishes (34) from (26) is that here in (34) is both token-reflexive and deictic, unlike here in the President's utterance, which is exclusively deictic.

Imagine that the caption under the photo from 1954 contained just the information that Jean Cocteau was the chairman of the jury at the Cannes film festival that year. Someone may then point to the photo and say either Here we see Jean Cocteau on the left or There we see Jean Cocteau on the left. When the singling out of the location of Jean Cocteau in the photo is done orally by means of here in combination with the description on the left, the reference to the space occupied by Jean Cocteau is purely deictic, not deictic as well as token-reflexive, which implies that here may be replaced by there. A caption like (34), however, invites the reader to interpret the image and the text as one unit from a cognitive point of view. The non-existence of the distal alternative there in such settings shows that intuitively, the indexical is token-reflexive. But the purpose of the written expression Here we see him on the left . . . is equally to cause the reader to find out where in the photo Jean Cocteau is. A sentence-initial here supports deictic processing of the indexical and causes the reader's immediate perceptual attention to the photo, and then the descriptive information on the left helps to uniquely pick out Jean Cocteau in the photo.

We believe the inability of a text writer to exploit deictic gesturing in the direction of the relevant space partly accounts for why there is impossible in (34). When a gesture is available as a procedural clue to identification of a space in front of the speaker (or possibly a space behind the speaker if the thumb is used to point backwards above the speaker's shoulder to something visible to the hearer that the speaker knows to be there), there is no need to indicate the hearer's proximity to the spatial referent by linguistic means, as the hearer is simply supposed to imagine a trajectory from the speaker's position to the position of the referent of the spatial indexical. The intended trajectory is iconically indicated by the speaker's gesture. In the absence of pointing, as in (34), the addressee has no procedural evidence available that enables him to anchor the distal there to an external stimulus. On the other hand, proximal here indicates that the reference of the indexical should be viewed as token-reflexive, as if the context visually represented in the photo that the caption comments on is one that the reader is included in.

Another illustration of how here may be used in captions is offered in (35), which is our English translation of a text appearing under a portrait of the King of Sweden in the Norwegian daily Dagbladet, 25 May 2011.

1. (35) HUNT: King Carl Gustav has allegedly taken part in ‘sex dinner parties’ with young women. Here it is pheasants he is chasing.

There are no pheasants in the image, just the King's face and his chest and a sporty hat suited to the occasion, the pheasant hunting event. A superficial glance at (35) and the photo it comments on may cause a reader to conclude that there is a lack of coherence between the first declarative and the next declarative starting with the indexical here, whose fronted position indicates that the reader is supposed to engage in a deictic processing of the indexical. It is in principle possible to test out an analysis of here in (35) as an anaphoric indexical, though the outcome of that pragmatic processing cannot possibly satisfy a reader's expectation of a relevant stimulus. If we replace here in (35) by there, the token-reflexive interpretation is ruled out, and an anaphoric reading of the indexical seems to be the only one available. Resolution of the designatum of an anaphorically used here in (35) as something like ‘during those dinner parties’ would be informed by antecedent information in the preceding sentence, and we would then arrive at the unlikely interpretation ‘The King is chasing pheasants while he is partying with young women’. The alternative processing of here in (35) as both token-reflexive and deictic stands a better chance of satisfying one's expectation of relevance. The King had for some time been in the media's searchlight because he was believed to have been chasing young women, in contrast to his chasing pheasants at the time when the photo session that included the photo appearing in Dagbladet was shot. Again, if the caption had not contained the Norwegian equivalent of the follow-up cleft sentence Here it is pheasants he is chasing, someone might point to the photo and say There he is chasing pheasants as an alternative to Here he is chasing pheasants, because the speaker's use of the spatial indexical would then be strictly deictic, not deictic and simultaneously token-reflexive as in the caption.

It appears that a similar combination of a token-reflexive and a deictic interpretation is necessary to resolve the reference of here in a very different kind of contextual frame. The written instruction Leave your tray here in a cafeteria invites the customer to look for the designated place, which may be a hole in the wall that the guest will discover after a brief visual search. The reference is deictic, because the guest is not supposed to leave the tray just anywhere close to the instruction, and it is token-reflexive because here could not have been replaced by there.

We conclude that if here is used in an act of deictic reference but cannot be replaced by its distal counterpart there without changing the import of the utterance or the grammatical well-formedness of the sentence, then its use is token-reflexive as well as deictic. It was shown in Section 2.2 that there are uses of here which cause the interpreter to adopt a token-reflexive and an anaphoric processing strategy that operate in tandem. The fact that the three uses are not mutually exclusive permits us to conclude that here has a unitary lexical meaning. By virtue of its procedural meaning, the indexical here encodes a relation of ‘spatial proximity’ between the source of information that enables an interpreter to resolve the reference of the indexical and the place where a token of it is produced, or the point in a written text where it appears. Resolution of the reference of here depends both on the hearer's knowledge of the lexical meaning of this word and on the interpreter's choice of a pragmatic processing strategy that takes contextual constraints on interpretation into account. Three strategies, the token-reflexive, the anaphoric and the deictic processing strategy, are available, and it is sometimes pertinent to combine two of them. A variety of procedural devices like the syntactic position of here, or its prosodic handling, or a gesturing that accompanies the indexical may facilitate the interpreter's choice of a single processing strategy in conformity with the communicator's derivational intention, but occasionally a combination of the token-reflexive and the deictic strategy is required.

It was shown that the information needed to identify a spatial referent by means of a deictic strategy may be acquired through hearing rather than vision. The chiming church bells is a source of information that the indexical here in an utterance of Here's the cathedral will direct the hearer's perceptual apparatus to. Here in that utterance is therefore deictic both for the speaker and the hearer. In contrast to this illustration, when someone says I am in here to an addressee who cannot see the speaker, the hearer's reference resolution process is based on his realization that the token-reflexive and the deictic processing strategy work in tandem. But what exactly do we mean when we say that here in I'm in here is both token-reflexive and deictic? Our answer is that it is token-reflexive for the speaker, and the hearer has to recognize this, but it is deictic for the hearer, who is observing which room the speaker's voice is coming from.

What we have claimed to be a combination of deixis and token-reflexivity in the written use of here in a caption that accompanies and comments on an image may be harder to account for. In regard to the Norwegian Jean Cocteau example in (34) above, we suggest that her ‘here’ is token-reflexive in the sense that the perceptual field indicating the frame of reference is limited to the image representing Jean Cocteau among a group of people participating in a parade in Cannes, but it is deictic in the sense that the indexical her in conjunction with the description ‘on the left’ directs the reader's visual perception to the location of a unique individual. The indexical is deictic because it does not refer indiscriminately to the space delimited by the pictorial frame but picks out a specific location. The King of Sweden example in (35) differs from (34) in that the King's face, the upper part of his body and his headgear dominate the image, so there is no need for a description of where the King is located. Here is token-reflexive in (35) because the frame of reference is limited to what can be observed in the image, but it is deictic for the reader of the newspaper, who is sure to read the caption and then look at the image. As the caption does not tie the King to a specified location, the reference of here is not de re but de dicto, to wherever the King had his portrait taken. The deictic aspect of the referential process is due to the fact that the pragmatic interpretation of the indexical involves visual perception.

3.3 Approximative modifiers of a deictic ‘here’ vs. a token-reflexive ‘here’

If the speaker adds the approximation marker somewhere to there, as in (36) below, the reference may still be deictic if the phrase there somewhere is accompanied by a gestural demonstration, and the statement will be true if the speaker is pointing in what is basically the right direction. Use of there in a deictic demonstration does not require a spatial referent that can be viewed or heard from the place of utterance.

1. (36) The cathedral is there somewhere.

Observe that there is no corresponding deictic use of the expression here somewhere. Here can only receive a token-reflexive interpretation in (37).

1. (37) The cathedral is here somewhere.

Somewhere suggests that the cathedral is not visible from the place of utterance. This is in order if the spatial term is distal there. One can point while uttering (36) and thus perform a successful act of deictic reference. We have argued that any deictic interpretation of here is constrained by what we called the sensory accessibility condition, so our prediction is that, when somewhere combines with here, the phrase here somewhere, in (37) above, does not satisfy this constraint on deictic use. While one may point in a particular direction while uttering (36), one will not point when uttering (37). This all adds up: the adverbial phrase here somewhere only permits the token-reflexive reading ‘somewhere around here’.

3.4 Differences between speaker's perspective and hearer's perspective

Imagine someone who is trying to locate a particular distant object by using binoculars. Having moved her gaze, through the binoculars, from right to left, she suddenly exclaims (38).

1. (38) Here it is!

The spatial referent is not within the hearer's perceptual range, and indeed, the communicated thought will not be fully relevant to the hearer before he is handed the binoculars so that he can look for himself and locate the object that is the referent of it. However, here – i.e. deictic here – is acceptable as long as the speaker, but not necessarily the hearer, is able to perceive the place where the object searched for is located.

The Norwegian lexicon contains an interesting, morphologically complex lexical item – her borte (‘here away’). Its function approximates that of the English expression over here. Given our view of the semantic properties and pragmatic consequences of a deictically used proximal spatial indexical, the existence of a lexical item like Norwegian her borte is not surprising. Her borte indicates either that the referent is at some distance from where the speaker is, due to borte, but visible, so that the accessibility condition on deictic use is met, or that the referent is at some distance from where the hearer is, but close to the speaker. While its lexical counterpart der borte (literally ‘there away’) may be used with reference to something beyond the speaker's perceptual range at the time of utterance, the contextual appropriateness of her borte in an utterance of (39) normally hinges on the speaker's visual perception of the escalators referred to.

1. (39)

The escalators may be seen from a vantage point shared by speaker and hearer, or the speaker of (39) may be located close to the escalators but at some distance from the hearer. In the latter situation, the speaker might be said to be taking the interlocutor's perspective, using the seemingly paradoxical combination of proximity and distance encoded in her borte in order to indicate that the referent can be perceived from the interlocutor's vantage point. The English expression over here has similar properties but the item over does not indicate distance as transparently as the Norwegian item borte ‘away, gone, disappeared’.

3.5 Here's Essien – deferred (metonymical) reference to the ball possessor, or spatial reference to where the ball is?

One use of here which is a bit tricky to classify may be characterized as a conventionalized feature of the speech of sports journalists’ reporting from televised football matches. There is a tendency for English-speaking football reporters to use an initial here in an utterance whose purpose is to point out which player possesses the ball at the moment. There is no deictic demonstration in such a case, here exemplified by (40), but observe that the condition for appropriate use of this conventionalized, genre-specific syntactic construction is that the player singled out by name must be the player in control of the ball.

1. (40) Here's Essien.

Thus, the visible location of the ball is, we think, a condition for legitimate use of deictic here in (40). Is the location of the ball also the referent of here? The addressee's, i.e. the viewer's, interpretation guided by the image of the TV screen enables him to form a representation of the Chelsea player Michael Essien from Ghana as the one who is in possession of the ball at the moment. The speaker's intention is not to inform the addressee that the player with the ball is called Essien. The information structure, and the content, is not the same as when someone utters This player is Essien or Essien is here. The explicitly communicated proposition may be paraphrased as ‘The ball is now with Essien’ or ‘The current ball possessor is Essien’. Thus, it looks as if the truth-conditional content has little in common with the syntactic form of (40); there can be no direct mapping from the encoded logical form of (40) to the proposition expressed. The proposition depicts the changed situation involving Essien as the new possessor of the ball. Here is not a constituent of the proposition of (40) qua locative adjunct, though it is significant to note Essien's position vis-à-vis the players of the opposing side and the players on his own side as well as the distance to the opponent's goal and the continuously changing distance between the ball and Essien's body as long as he remains in control of the ball.

What if we imagine that the real semantic value of here in (40) is the current ball possessor rather than a location? If this is our analysis, the use of here in (40) is a case of deferred reference (Nunberg Reference Nunberg1979, Reference Nunberg1995, Reference Nunberg, Reimer and Bezuidenhout2004a, Reference Nunberg, Horn and Wardb), because the deictic indexical points the addressee to where the ball is, and the ball is metonymically associated with the possessor of the ball referred to by means of his name. The current location of the ball might be analyzed as Nunberg's index (1995) of the speaker's act of deferred reference to Essien as the man with the ball. Here would then draw the TV viewers’ attention to this visible index, which serves as a metonym for the intended referent, the attributively understood entity ‘the current possessor of the ball’.

On the analysis outlined in the preceding paragraph, there is a very considerable gap between the syntactic form and sentence meaning of (40) and the proposition that an utterance of (40) communicates. A competing analysis that reduces this gap makes the true referent of here the location of the football at the time of utterance. The communicated information would then be that the location of the ball and the location of Essien coincide at the time of utterance. Here refers to the former and the noun phrase Essien to the latter. The utterance tells us about the relationship between the football and a specific player. Essien is where the ball is, which implies that Essien is the current possessor of the ball. This analysis does not involve a deferred interpretation of here as a necessary component. That the locations of Essien and the football coincide is probably not literally true, because the ball possessor is only intermittently where the ball is, just when there is physical contact between the player and the ball. Still, we have a sufficient number of indicators of what inferential processes will make the grammatical form of (40) relevant. This analysis is arguably simpler than its competitor and less uncongenial relative to more standard thinking about the relationship between linguistic meaning and communicated thought/proposition – which may be considered a good enough reason to adopt it.

Viewers of a televised football match will automatically follow the ball with their eyes and observe who is in possession of the ball at any moment, just like the reporter who produced (40). An utterance of (40) will not be relevant to the audience unless it is meant to describe the relationship between the location of the ball and the particular player who is in control of the ball, or who is passing the ball, at the time of utterance. This means that here either refers metonymically to the ball possessor or to where the ball is, the latter being our preferred analysis.

3.6 Summing up

Based on what we believe our examples of the strictly deictic use of here in the present section to have in common, we now propose to define the special conditions pertaining to deictic here in the following way: The interlocutors’ perspective on the referent is projected from the speaker's here-and-now to a deictically singled out, designated location within the speaker's perceptual (most frequently visual) field. The location referred to does not coincide with the spot occupied by the communicator but is projected to a place close enough to be seen, or otherwise perceived, from the spatiotemporal vantage point of the speaker. Since the deictic demonstration is normally performed for the benefit of the addressee, the location to be identified as the referent of here will typically also be within the addressee's perceptual field. While perceptibility as a prerequisite may be said to apply only to the speaker's ability to see or hear the place referred to, some of the stimuli presented in this section are clearly irrelevant unless the addressee can also perceive the spatial referent. However, stimulus relevance should be judged in relation to the speaker's belief about what information is accessible to the addressee, so when the addressee of an utterance of I am in here fails to hear that an utterance is being directed to him, the utterance will still be relevant if the speaker is justified in believing that he was speaking loud enough and clearly enough for the delivered message to be within the addressee's perceptual range (see Sperber & Wilson Reference Sperber and Wilson1986).

We hope that the content of Section 3.2 has justified our claim that a token of here may be intended to be processed pragmatically as a deictic indexical whose reference must be resolved with the help of the interpreter's perception of the location of a designated object but at the same time as a token-reflexive indexical because it refers reflexively to the location of the perceptual target, whether this is the location of an object represented in a picture or the location of a person whose voice indicates where he is at the moment of utterance.

4.1 The doctor's ‘here’ and the patient's ‘there’

While several philosophers of language have cited the pointing-at-a-place-on-a-map phenomenon as a case where here is a demonstrative rather than a ‘pure’ indexical, some linguists have cited an invented conversation between a doctor and her patient as an interesting counterexample to the hypothesis that ‘proximal’ is invariably to be understood as close to the speaker, an attribute that the referent of both medial and distal demonstratives is supposed to lack. This invented doctor–patient conversation caught our interest and was an important impetus for the writing of this paper.

The scenario is this: a GP has started palpating points on her patient's body, and is asking the question Does it hurt here? While the patient could answer affirmatively by simply saying Yes or negatively by saying No, it is also possible for him to produce a full sentence, as shown in (41). The English-speaking doctor will then have a choice between here and there but the patient will use distal there, or so the argument goes, even though the referent is a part of the patient's own body.

1. (41)

We hold that the important thing to notice if one wants to understand the principle that governs the distribution of here and there in (41) is that the person who introduces the discourse referent represented by the spatial indexical is the one who defines the target of the referential act. Psychologically the patient as a sentient being is not a legitimate referent, as the domain of reference is restricted to the patient's body, or maybe more narrowly to his torso. The patient's indexical does not assume the role of a token-reflexive term, because the doctor has decided that the domain of reference is the body examined, and the reference is therefore restricted to the point on the body which is deictically singled out by palpation.

Janssen (Reference Janssen, Geiger and Rudzka-Ostyn1993) observes that Dutch spatial indexicals operate in the same way as the English ones. One likely version of the Dutch conversation is seen in Janssen's illustration presented in (42).

1. (42)

Janssen's comment is that, ‘the differentiation between the domain of hier/deze “here/this” and daar/die “there/that” cannot be characterized as “in the vicinity of the speaker” and “not in the vicinity of the speaker”’ (Janssen Reference Janssen, Geiger and Rudzka-Ostyn1993:768, quoting Lyons Reference Lyons1968:275).

Janssen lists the following three features as the only factors relevant to the understanding of the distribution of hier ‘here’ and daar ‘there’ in an exchange like (42): (i) the speaker's vantage point, (ii) the salience of the place in question with regard to the speaker's vantage point, and (iii) the speaker's referential concern for the salient place with regard to his vantage point (Janssen Reference Janssen, Geiger and Rudzka-Ostyn1993:768). These three points are believed by Janssen to account for not only the data above but also the difference between licit use of hier and daar used anaphorically, as in (43), where the Dutch narrator is said to present the event from the point of view of Fred's parents in Montreal, compared to (44), where the narrator ‘sees the event at hand from his actual situation outside of Montreal’ (Janssen Reference Janssen, Geiger and Rudzka-Ostyn1993:770).

1. (43) Dutch

2. Fred arriveerde in Montreal. Hier hadden zijn ouders al lang naar hem uitgezien.

3. ‘Fred arrived in Montreal. His parents had long been awaiting him here.’

1. (44) Dutch

2. Fred ging naar Montreal. Daar moest hij zijn voor een conferentie.

3. ‘Fred went to Montreal. He had to attend a business meeting there.’

Janssen's three points cited above may shed some light on the distribution of the proximal and the distal spatial indexicals when they are used as anaphors, but we believe the difference between the doctor's deictic here and the patient's anaphoric there to require a somewhat different handling. In our opinion the presence vs. absence of a manual behaviour accurate enough to uniquely determine the referent is the most important parameter for an adequate analysis of (41)/(42) and similar data from other languages.

What needs to be accounted for in (41)/(42) is the doctor's ability to use the proximal spatial indexical without violating any principle governing the use of this item, and also, more remarkably, the constraint that rules out the patient's use of the same indexical, here/hier, even if the referent is a point on the patient's own body, and even if the referent introduced by the doctor is a maximally salient entity at the time when the patient produces his answer. The indexical is not used token-reflexively by the patient in such a case. Indeed, the fact that the proximal alternative is very unnatural, if not prohibited (in Dutch as well as English) shows that the patient's reference to a point on his body has nothing to do with token reflexivity. The domain of reference was determined by the doctor activating the spatial referent for the patient. If the patient is able to see the place referred to, he will identify it on the basis of two simultaneous and independent sensory impressions, a visual stimulus and a tactile stimulus; if the palpated point is somewhere on the patient's back, the tactile sensation is the only source of information leading to the patient's resolution of the reference of the doctor's here. The fact that the reference is only made accessible to the patient via perception makes it deictic, even if the patient, as interpreter, is unable to see the place referred to. Although the palpated spot belongs to the patient's body, we do not think of the place singled out by the doctor as the place where the patient is located at the time of utterance. After all, that spot is an inalienable part of this person wherever he goes. For this reason the patient cannot be using here as a token-reflexive indexical in his response to the doctor.

By analogy with a number of illustrations offered in Section 3, we notice that it is possible to shift the reference from the spatiotemporal centre of the conversation taking place in the doctor's office to a precisely delimited domain of reference. This is what the doctor achieved by letting his proximal term be accompanied by an extra-linguistic demonstration directed to a point on the patient's body. In his act of referring to the same place that the doctor's here referred to, the patient is not engaged in deictic reference. Because that place was already introduced as a discourse referent, any ensuing reference to the same place is a case of anaphoric use of the spatial indexical. On a more general note, the antecedent that a hearer must activate in order to resolve the reference of an anaphoric demonstrative will not necessarily be a proper name or a descriptive noun phrase. The linguistic antecedent may be an indexical just like the later anaphor. An indexical can be anaphoric upon an antecedent indexical whose referential interpretation is perceptually grounded, and the deictically used proximal indexical in the doctor's utterance serves the role of antecedent for the anaphoric, non-proximal demonstrative in the patient's answer.

We hold, with Janssen, that it is not the nearness of the doctor to the referent that allows her to use English here or Dutch hier, but unlike Janssen we believe that the speaker's perception of the spatial referent is a sufficient condition for deictic use of here/hier accompanied by ostensive extra-linguistic indication of the spatial reference. As long as the patient is not himself touching the object first picked out for reference by the doctor, the patient is not engaged in an act of deictic reference. He is then using his spatial indexical for anaphoric reference to the spot referred to by the doctor, and this, we hold, is why the patient will use the distal member there.

Use of proximal indexicals across participants in a conversation appears to be problematic in many languages, and this is not only true of spatial indexicals. As Robin Lakoff says (Lakoff Reference Lakoff1974:349), ‘[t]his may be used only if the two sentences are uttered by the same speaker’, and ‘that can be used by a speaker to comment on an immediately prior remark by another’. There is a speaker activation constraint on proximals which causes a responding hearer to avoid a proximal indexical like here when it is anaphoric upon the first speaker's deictic reference with the proximal member of the pair here/there, and to use the corresponding distal term there instead, not because the distal term necessarily signifies a geographical distance between speaker and referent but because distal there is the neutral member of the pair.

The speaker activation constraint on appropriate use of the proximal term obviously does not apply to token-reflexive uses. A question like Do you like it here? cannot be answered by #Yes, I like it there, even though Does it hurt here? should be answered by Yes, it hurts there when no gesture accompanies the indexical.

We believe the difference between the two pragmatically distinct phenomena of deixis and anaphora to offer us an explanation for the distribution of the spatial indexicals in (41)/(42). The doctor's here stands for ‘the place that I now single out (by deictic demonstration)’, but the patient's there stands for ‘the place that you have singled out (by deictic demonstration)’. Although the doctor's referent and the patient's referent are the same spatial entity and the two different modes of presentation of the reference is truth-conditionally irrelevant, the two indexicals here and there are cognitively distinct items. They represent distinct psychological modes of presentation in Recanati's sense (1993), and this difference is reflected directly by the doctor's choice of the proximal term for deictic reference and the patient's choice of the distal term for anaphoric reference to the object antecedently picked out by the doctor.

The conditions governing the use of here are not satisfied in the patient's case. This generalization holds at least for those languages that respect the speaker-activation condition on the anaphoric use of proximal demonstratives. It turns out, however, that not all languages behave like English and Dutch with regard to the distribution of spatial indexicals in the doctor–patient dialogue, and we are interested in this difference between languages that obey the speaker-activation condition and those that do not. Russian (Gundel et al. Reference Gundel, Hedberg and Zacharski1993) is claimed to belong to the latter group of languages. In Russian, one speaker may use a proximal term which is then echoed by a different speaker's use of the same item in a subsequent utterance to indicate coreference with the first speaker's indexical. With regard to the doctor–patient dialogue, we hold that an English patient's use of a distal spatial indexical and a Russian patient's use of a proximal spatial indexical under the same contextual conditions should not cause us to conclude that the speaker of English treats the indexical as anaphoric upon the doctor's previous use of the proximal option and activation of the selected spot as discourse referent, while the speaker of Russian chooses a production strategy which implies that the patient's indexical does not qualify as anaphoric, thus permitting him to use the same proximal item as the doctor. Rather, the patient's indexical is anaphoric in both languages, but while the speaker-activation condition on the use of proximals applies to the English language, it is much less certain that it applies to Russian (see the discussion in Section 4.3.2 below).

4.2 Looking at maps

The following illustration shares certain important elements with the doctor–patient dialogue but differs from it in some interesting respects. Suppose two individuals A and B are looking for the city of Ravenna on a map of Italy. A discovers where Ravenna is, points to where the name ‘Ravenna’ appears on the map, and utters Here it is, alternatively Ravenna is HERE, which causes B to react as shown in (45).

1. (45) A: Here it is.

2. B: I didn't know it was there.

Speaker A could have used the alternative forms There it is or Ravenna is there, but #I didn't know it was here or #Is it really here? would not be an option for B in (45). A's use of here is based on a projection from the point where her fingertip is to a mini-domain of reference restricted to the opened map on a table in front of the interlocutors. Speaker B will use there, because for B this is anaphoric reference to something antecedently referred to by A. One difference between (45) and the conversational exchange between the doctor and the patient is that the patient was as close to the spatial referent as he could possibly be, while person B in (45) does not have to be as near to the referent as A. Provided B is able to see where on the map of Italy A's index finger is placed, it is irrelevant how close to the map or far from the map B's eyes are at the time of utterance. As one reviewer pointed out, A and B could be equidistant from the map with both of them looking at it, or B could even be holding the map. The reason why the proximal indexical here is not an option for B in the context of A's previous utterance is that it would have violated the speaker-activation condition on anaphoric use of proximal indexicals.

It is possible to use the proximal term here even if one is looking at a map without an accompanying gestural demonstration, but uttering (45) with no manual gesture would not be an act of telling the interlocutor where Ravenna is. It would rather be equivalent to an act of A's telling B that A has located Ravenna on the map, an alternative to an utterance of I've found it. If one's intention is to tell someone where in Italy Ravenna is, one has to make sure that the hearer can see the map of Italy and see the point that his or her index finger is pointing to or touching. The proximal term here of Here it is! is motivated by the fact that the speaker is looking directly at the name ‘Ravenna’ discovered a moment earlier. There it is! would have expressed the same proposition as Here it is!, because deictic there is the neutral term that may be used freely even in contexts where deictic here is permitted. Even though B can see the point on the map which A is referring to, B is not free to choose between here and there as spatial indexicals in (45), because A is the speaker who picked out Ravenna on the map. B's item there is coreferential with A's item here, which we identify as the linguistic antecedent of B's later anaphoric indexical. We call A's here the antecedent of B's there, even though here is not a descriptive phrase but is itself an indexical. The information needed to resolve the reference of B's there is not information encoded in A's antecedent here, obviously, because here is a deictically used indexical whose referent can only be resolved through an act of interpretation grounded in visual perception.Footnote ⁷

Observe that in uttering (46), a speaker who is pointing to the name Ravenna on a map of Italy may perform a relevant utterance even if the hearer knows that he is in Ravenna. In fact, it is rather unlikely that the hearer does not know which city he is in.

1. (46) We're here.

On its most natural interpretation, the meaning of the utterance of (46) produced by someone who lets her index finger pick out the name Ravenna on the map is not ‘We are in Ravenna’. Rather, the reference of here in (46) will be determined in terms of where in Italy the city is located relative to certain other Italian cities or regions that the hearer may already be familiar with. The meaning of (46) is that the interlocutors are where it says Ravenna on the map. Although here is an indexical that encodes no description of the referent, it does not refer de re to Ravenna in (46) uttered in the specified context. Rather, the referent is ‘where my forefinger is touching the map’.

It is possible, however, that the hearer does not know the name of the city they have arrived in. In that kind of context, the indexical here in an utterance of (46) has a dual function. On the one hand the utterance informs the hearer that Ravenna is the name of the city they are in, because it says Ravenna where the tip of the speaker's forefinger is; on the other hand the utterance of (46) informs the hearer of the location of Ravenna on a map of Italy.

4.3 A cross-linguistic examination of deictic and non-deictic (anaphoric) uses of spatial indexicals