The honorific third person plural in Slavic

Although much has been written about polite forms of address in Slavic, the grammatical expression of respect for a person that does not take part in the conversation has hitherto received little to no attention. Yet this type of honorific reference is widely found in the Slavic languages, especially in dialects and substandard varieties. In the present article I shall distinguish between three types of honorific reference to a person spoken about and focus on the type that I think is most current in Slavic dialects, viz. honorific reference to elder relatives, and, by extension, other familiar older persons and priests. I shall first present an analysis of the use of the construction and the kind of respect involved. One of the characteristics that make this honorific plural interesting from a typological point of view is that it is primarily used for respected persons the speaker has an in-group relationship with: if there is too much ‘horizontal distance’ it is not used. Then I shall give a picture of the distribution of this type of honorific third person plural across the Slavic language area. It will become clear that it is found across a more or less continuous area from Slovene in the southwest to central Ukrainian in the northeast.


Introduction
The phenomenon that will be discussed in this article drew my attention a couple of years ago when I was working on the Croatian dialects of the Burgenland and its surroundings. 1 It is the use of the third person plural when referring to a single person who is respected by the speaker: 2 (1) Njeguov otadz už vig živu, sat su sedamdesiet lit stari. 3 live.pres.3pl; be.pres.3pl; old.masc.pl 'His father is still alive, he is now seventy years old. ' (Petrovo Selo / Szentpéterfa; my material) (2) Ja vidim oš tetu Linu Honimonku kako su s auta skočili [. . .]. Oni su bili dovica [. . .]. be.pres.3pl; jump.l-part.masc.pl; they.nom.masc; be.pres.3pl; be.l-part.masc.pl; 'I still see [= lively remember]  The respected person who is referred to by the third person plural is not the person spoken to and does not have to be present. 4 In Houtzagers (2012) I gave a picture of where in the Burgenland the construction can be found and presented an analysis of its usage: what persons it can refer to, what kind of respect is involved, etc. I shall summarize the results that are relevant for this article in Sects. 2 and 3.
Soon afterwards I started to look for examples of this honorific plural in other Slavic language varieties, first in those spoken in the neighborhood of Burgenland Croatian, but I soon expanded my area of research, as it became clear that it could be found in a considerable part of the Slavic linguistic area, more often in dialects and substandard variants than in the literary languages, and mostly in the speech of the oldest generations. This construction in Slavic gives rise to a number of questions: (a) What are the morphosyntactic and semantic / pragmatic characteristics of the construction? What are the similarities and differences between this construction and other constructions in Slavic that express respect for a third person? (b) How special is the construction from a typological point of view? (c) Where in the Slavic-speaking area can this construction be found? (d) Can we say anything about the origins of the construction? Where, when and how did it arise and how can we explain that it covers the area it does nowadays?
In the present article I shall focus on (a), (b) and (c). I intend to come back to the questions listed under (d) in future publications.

The construction in more detail
I shall call the construction presented above 'HON3pl-A', where 'HON3pl' stands for 'honorific third person plural'. The letter 'A' distinguishes this construction from two similar constructions ('HON3pl-B' and 'HON3pl-C') that will be introduced below, in Sects. 4 and 5. This article focuses on HON3pl-A. For reasons of commodity I have introduced the abbreviations 'S' for 'speaker' and 'R' for 'referent', that is the person referred to / spoken about and respected by S. The HON3pl-A is more than merely a syntactic construction in which a subject in the singular agrees with a predicate in the plural, as in (1) and the first sentence of (2). Example (3) and the second sentence of (2) show that this 'peculiar' agreement does not have to be present: when R is represented by a pronoun, nothing special happens on the syntactic level. What happens there is special from a semantic or pragmatic point of view: a plural pronoun refers to a single person. This plural pronoun is not necessarily the subject of the sentence and it does not have to be a personal pronoun. From example (3) above it can be seen that the plural pronoun can be in another case than the nominative. Example (4) below shows R represented by a plural possessive pronoun: (4) Diede Jure Jurdić su na grozlje trganje njihovoga vnuka Jozu pozvali (Joza danas živi na sievernoj Moravi). be.pres.3pl; their.masc.acc.sg; call.l-part.masc.pl 'Grandfather Jure Jurdić invited his grandson Joza for the wine harvest (Joza today lives in northern Moravia).' (Moravian Croats; Lawitschka 2005, p. 86) As far as I know, in those language varieties that distinguish between genders in the plural (as Croatian does in the l-participle and other participles, the adjective and the personal pronoun), the HON3pl-A takes the masculine form, even when R is female (see example (2) above).

Usage of HON3pl-A in the Burgenland. 5 Typological remarks
In Houtzagers (2012) I studied the usage of HON3pl-A in Burgenland Croatian (in those dialects in which it occurs) in detail and came to the following conclusion: If we attempt to formulate an invariant for the class of persons referred to by the honorific plural under discussion, my proposal would be: 'persons who are respected by the speaker in the affectionate way a child respects (or is by tradition bound to respect) an elder relative'. This formulation would, in my opinion, include the respect and affection for a grownup who was good to him and whom he felt close to as a child and the respect that a believer in a traditional rural community has for a priest and for God. 6 If nonrelatives are referred to there is some freedom as to whether or not the speaker uses the honorific construction. Lack of closeness or personal acquaintance does not always prevent the use of the construction but can apparently be 'overruled' by the feeling of respect. (Houtzagers 2012, p. 298).
In the overwhelming majority of examples of the HON3pl-A in my Burgenland material, the Rs were elder relatives. Very often they were no longer alive, which stands to reason because most of the informants had themselves reached an advanced age at the time the relevant utterances were recorded. In almost all cases in which the Rs were not elder relatives, they were grownups whom S knew when s/he was a child, or priests. In exceptional cases R could be a respected older person whom S did not know as a child or did not know personally at all. 'Closeness' is an important indicator for the use of the HON3pl-A. With very few exceptions, R is personally known to S and shares relevant domains of his life with S: in most cases S and R are part of the same village community, in less frequent cases S and R have been friends or close colleagues for a long time without S and R belonging to the same village community.
In theoretical works on politeness, the decision which politeness strategies to use in a specific utterance is, in general, based on three factors. In Leech's terminology (2014, pp. 10-11, 84, 103, 106-109) these factors are: (a) vertical distance; other authors often call this factor 'power' (Brown and Gilman 1960;Brown and Levinson 1987) or 'status' (Haase 1994, pp. 21-22); (b) horizontal distance; often simply called 'distance' by other authors (Brown and Levinson 1987;Haase 1994, pp. 21-22); Brown and Gilman (1960) use the term 'solidarity' for horizontal closeness; (c) weigthiness of the transaction; in other words: how much and what kind of politeness is needed on the specific occasion on which the utterance takes place?
For the present description of HON3pl-A we can do without factor (c). 7 As can be seen above, the terms used to designate factors (a) and (b) differ among authors. However, what is meant by the respective terms largely coincides. Especially in the case of factor (a) I prefer Leech's term over those proposed by others, and for the reason Leech gives (2014, p. 106): the term 'power' would suggest that social superiority of R over S is relevant, which it is not. Different cultures give different significance to the factors potentially involved in vertical distance. In some cultures the age factor is more important than in others, and for the HON3pl-A in Slavic age seems to be the only relevant factor on the vertical axis. Once we have chosen the term 'vertical distance' for factor (a), 'horizontal distance' is the perfect term for factor (b). 8 Factors (a) and (b) above can be viewed as axes in a two-dimensional space. A combination of certain positions on these axes is indicative for the use of the honorific construction under discussion: the vertical distance must be not too small, the horizontal distance must be not too great.
For the HON3pl-A in Burgenland Croatian, the vertical axis stands for the variable 'age'. The vertical distance required to use HON3pl-A is 'at least one generation'. For the vertical axis it is relevant who is highest: R is above (i.e. older than) S. The vertical axis is not gradable in the sense that the greater the vertical distance, the greater the need to use HON3pl-A. A distance of one generation is necessary and sufficient. Grandparents are not more likely to be referred to by HON3pl-A than parents.
Horizontal distance is more complicated, because it consists of two components, between which the hierarchy varies. Probably S has some freedom of choice here. The HON3pl-A is used if (i) S and R have an in-group relationship; this usually means that they are personally acqainted. The strongest in-group relationship is being family. Another (less strong) ingroup relationship is being part of the same village community; (ii) S likes or respects, or at least does not dislike or disrespect R.
The combination of closeness (often in the form of a family relationship) and liking and respect gives HON3pl-A the special flavour that I have tried to describe above ("the affectionate way a child respects (or is by tradition bound to respect) an elder relative", Houtzagers 2012, p. 298) and what Berger (1996, p. 31) calls "respektvolle Vertrautheit" 'respectful intimacy'. Very often HON3pl-A occurs in combination with endearing forms referring to elder family members, e.g. mamulka, tatulek (endearing diminutives for 'mother ', 'father', respectively, cf. Nitsch 1968, p. 66). 9 It is not the same kind of respect that is based on social distance. On the contrary, too much social distance is contraindicative for the use of HON3pl-A. I shall give some examples of varying hierarchical relationships between (i) and (ii): 10 eral, I do not fully agree with Leech, because a speaker can decide to be more or less 'polite' depending on what s/he wants to attain. For the present account however, especially since it deals with reference to a third person, factor (c) indeed seems less relevant. 8 Head (1978) works with a variable 'respect or social distance', which is not defined and probably represents a single dimension. Especially Leech's term for the vertical axis, viz. 'vertical distance', is to be preferred over 'power' or 'status': Rs with much power and status are most likely not to be referred to using the HON3pl-A in view of the horizontal distance to S (see below in the present section). 9 One of my informants on Slovak told me that her mother (born in 1953 in a village not too far from Bratislava) sometimes used an endearing diminutive with HON3pl-A when referring to the parish priest: Farárko boli [. . .] 'The priest (diminutive) were [. . .]'. 10 The observations described in the bulleted list are mainly based on two sources (Lawitschka 2005 andPokorný 1992) which show ample use of the HON3pl-A and in which use of the HON3pl-A seems not to be optional. In many Slavic dialects, however, the HON3pl-A can or cannot be used, probably because the HON3pl is gradually disappearing there (or was disappearing at the time when the relevant forms were attested). In such cases it would only be possible to say anything about such subtleties as varying hierarchical • My material contains only one instance of an elder family member not being referred to using the HON3pl-A (an uncle who physically abused S's mother), which suggests that being an elder family member is enough to be referred to by means of the HON3pl-A, and is enough to guarantee the respect needed for the use of HON3pl-A (except in extreme cases). If that is true, horizontal component (i), if strong enough, overrules (ii). 11 • I do have examples of older persons (not family) who are respected and liked by S, but have no in-group relationship with him or her and therefore are referred to by singular forms. Here too, horizontal component (i) dominates. • There are (rare) examples of Rs referred to using the HON3pl-A who are not members of the S's family, with whom S has no personal acquaintance or who even had died long before S was born. Apparently, in cases such as these, the respect is so great that component (ii) overrules component (i).
School teachers, doctors, etc. are usually not referred to with HON3pl-A. On the vertical axis they can qualify for HON3pl-A if they are at least one generation older than S (not because of their social status, which is not relevant on the vertical axis), but on the horizontal axis they are usually excluded by absence of an in-group relationship, which can only be compensated by great affection felt by S. What is interesting from a typological point of view is that horizontal distance influences the decision whether or not to use HON3pl-A in a non-standard way if one compares it with what is reported in general works about politeness in languages. In languages in general, as a rule, the greater the horizontal distance, the greater the chance of honorific forms (Brown and Levinson 1987, pp. 74-78). For HON3pl-A the opposite holds: the smaller the horizontal distance, the greater the chance of honorific forms. 12 This section was about the usage of HON3pl-A in Burgenland Croatian, about which I have ample material. In the rest of this article, the reader will see that HON3pl-A in Slavic is by no means limited to Burgenland Croatian. Of course I did not study the use of the construction as thoroughly for each place where I found it and I cannot be sure that everywhere in Slavic where HON3pl-A exists, it is used in precisely the same way as described here. However, the set of possible Rs of HON3pl-A is specific enough to distinguish between HON3pl-A on the one hand and HON3pl-B or -C (introduced in Sects. 4 and 5) on the other. For instance, if an informant refers to his deceased mother with a plural form, we can be confident that we have to do with HON3pl-A.
As we shall see at the end of Sect. 5 and in Sect. 7 below, there are Slavic language varieties where we have to assume a honorific third person plural with a wider set of Rs than presented in the present section. Such cases will be analyzed as combinations of HON3pl-A and HON3pl-C. relations between factors (i) and (ii) if much material would be available for each individual dialect, which is never the case. 11 Probably the use of the HON3pl-A when talking about a priest must be explained along the same lines. As far as honorification is concerned, a priest is put on a par with a father (not because of his age but because of his moral authority) and is therefore (morally) a generation older on the vertical axis and family on the horizontal axis. 12 It is possible that 'having a comparable social status' is also an in-group quality, but here I hesitate for two reasons: (i) in my examples in which S and R had a different social status, R also did not belong to the same village community as S and / or was not liked by S, so that I cannot be sure that social status is relevant at all; (ii) social status is normally associated with the vertical axis. The reason for me to consider it as a factor relevant for horizontal and not vertical distance is that the direction of the scale is the same as with the other components of the horizontal axis: the greater the distance, the smaller the chance of HON3pl-A.

HON3pl-B
In Sect. 1 I wrote that the R of a HON3pl-A construction does not have to be present to the conversation. Honorification specifically meant for an R who is not spoken to but who is present to the conversation (henceforth 'HON3pl-B') also exists in Slavic, but it is a separate phenomenon that should not be confused with HON3pl-A. HON3pl-B can occur in the same dialects that have HON3pl-A and in the instances that I attested it was expressed by the same grammatical means as HON3pl-A. However, the set of Rs the construction refers to is not necessarily the same. In the village of Koljnof near Sopron an informant, who was at least 30 years older than me, said to his wife, referring to me: (5) Oni se po vedešinsku bolje znadu. they; know.pres.3pl 'He knows the dialect of Vedešin better [than ours].' From this example it is clear that HON3pl-A and HON3pl-B are different phenomena. Neither on the vertical nor on the horizontal axis, the R in (5) (the present author) would qualify as an R for HON3pl-A. 13 The dialect of Koljnof also has HON3pl-A: 14 (6) Gospodin su vedešinac (my material). be.pres.3pl 'The priest is from Vedešin.'

HON3pl-C
Russian has a construction that, again, is very similar to HON3pl-A grammatically, but differs with regard to the set of Rs, the kind of respect involved and stylistic connotations. I shall call this type of honorification 'HON3pl-C'. An example: In this construction, the vertical axis is clearly a matter of social status (R significantly higher than S) and on the horizontal axis in-group relationship seems irrelevant or even contraindicative. In present-day Russian the use of HON3pl-C almost always has a comical effect. The reason for this is that it refers to a culture of days gone by, when persons with a lower status had to show explicit respect to persons with a higher status, even in their absence. Because of the high social status of R and the significantly lower status of S implied, in present-day Russian HON3pl-C is almost always used ironically. 15,16,17 I know of no publication in which HON3pl-C in Russian, let alone its origin, is discussed at some length. 18 The same construction can be found in 19th century German (rarely in Modern German), when the sentence subject contains a prestigious title (e.g. der Herr Graf, literally 'Mr. Count') and S is a subordinate or someone in a position of 'servitude', e.g. a waiter, a coach driver, etc. In example (8), from a 1867 novel by Theodor Storm, the narrator asks an inn-keeper the whereabouts of his friend. The inn-keeper answers: (8) [. . .] der Herr Professor sind mir wohl bekannt; sie haben zu Anfang ihres hiesigen Aufenthalts ein Vierteljahr in meinem Hotel zu Mittag gespeist. (Storm 1987, p. 22) be.pres.3pl; they; have.pres.3pl; their.gen.masc.sg 'I do know the professor. At the beginning of his stay here he took his luncheon in my hotel for a quarter of a year.' In German, too, ironic connotations are possible. I know only one publication in which HON3pl-C in German is discussed, viz. Findreng (1976, pp. 53-54, 95-106). Findreng gives no information about the origin of the construction, from what period onward one can find it in German texts and its distribution in the German-speaking area. Another question that comes to mind is whether German HON3pl-C could have served as a model for Russian HON3pl-C. I think this certainly is a possibility, but separate research would be needed to answer this question. It would have to be made plausible that in some relevant period there was sufficient German linguistic influence in Russia for a replication of this type to take place.
In this article I treat HON3pl-A and HON3pl-C as similar but different, although often cooccurring phenomena. This is because in most Slavic dialects in which I found HON3pl-A, I found no examples of socially higher placed persons being referred to using a third person plural, and, reversely, in the German and Russian speech varieties discussed above it does not seem possible to refer to one's mother using a honorific third person plural. 19 However, as we shall see in Sect. 7 below, there are Slavic dialects (especially Kajkavian, Slovene and Slovak) which seem to show both HON3pl-A and HON3pl-C. It could be argued that in such cases we would have to assume a HON3pl-D, the set of Rs of which would equal the union of the sets of Rs of HON3pl-A and HON3pl-C. Yet I have chosen not to introduce a HON3pl-D. The reason for this is that in Slavic, honorific reference to elder relatives and priests seems to be more general and more stable than to other respected persons. There are many locations for which examples of the use of the HON3pl-A were found when the R was an elder relative or a priest and examples of the non-use of the HON3pl-A when the R was another socially higher placed person. The reverse situation, although rare, has also been found. This seems to testify that honorific reference to other respected persons is a separate phenomenon that, if it is present, usually coexists but does not have to coexist with honorific reference to elder relatives and priests.

Earlier work on the subject
There exists a substantial amount of linguistic literature about politeness, forms of address, etc. in Slavic languages (see the references given in Betsch 2007;Betsch and Berger 2009), but little attention has been paid to the linguistic expression of respect for a person who is not the addressee. In publications in which HON3pl-A, -B or -C are mentioned explicitly, the scope is usually limited to a regional variant of a language or one language as a whole.
As far as I know, the only publications so far in which HON3pl-A or HON3pl-C is treated as a phenomenon shared by several Slavic languages are Berger ( , 1996 and Corbett (1983Corbett ( , 2009). Berger (1995, pp. 42-43) briefly discusses "die pluralische Kongruenz des Verbs beim Bezug auf nicht anwesende Personen, die einen höheren Status haben" 'the plural verbal agreement with subjects that have a higher status than the speaker and are not present' in Slovak, Russian, Polish, Ukrainian and Belarusian. Berger (1996, pp. 27-31) is probably the most extensive discussion of HON3pl-A and HON3pl-C available. The author discusses (and distinguishes) HON3pl-A in Polish, Czech, Slovak and Russian and HON3pl-C in Russian. With regard to the class of Rs of HON3pl-A and the relationship between S and those Rs, he largely agrees with my observations about Burgenland Croatian in Sect. 1 above ("respektvolle Vertrautheit", 1996, p. 31). Corbett (2009, pp. 346-347) donates only half a page to the subject, but he gives the most complete list of languages in which the construction can be found: Polish, Czech, Slovak, Slovene, Kajkavian, Belarusian, Ukrainian and Russian. The same languages (except Kajkavian) are already mentioned in Corbett (1983, p. 41 fn. 8), together with sources for the various languages.

The distribution of HON3pl-A in Slavic
In the following I shall concentrate on the third question formulated in Sect. 1: "Where in the Slavic-speaking area is this construction found?" I have gathered information from all Slavic languages and I have tried to establish the distribution of the HON3pl-A as accurately as Apart from instances in dialect texts, I have tried to find observations made by linguists about the various languages, not only the dialects but also the standard and substandard variants. I have also consulted native speakers of and experts on the various languages (see footnote on first page). Plus and minus signs on the maps stand for observations on presence or absence of HON3pl-A over a larger area. These signs will always be commented upon in the text.
For some language areas it is easier and less time-consuming to find a reasonable number of examples with a satisfactory geographic distribution than for others. This explains the differences in coverage in the report of my findings below. The ideal situation is when there is a (not too small) collection of geographically well spread dialect texts available for a given language in which the speakers give accounts of their own lives, such as Michálková (1976) for Czech or Padlužny (1990) for Belarusian. 21 But the situation is not always ideal and therefore, for some (parts of) language areas one has to consult separate dialect monographs and hope that they contain useful texts. This, for example, is the case for large parts of the Bosnian-Croatian-Serbian and Slovene language areas. For these languages a fuller coverage would in principle be possible, but would require additional time investment.
Because there is a good chance that there is a connection between the system of honorific reference to a third person and that of honorific reference to the addressee in a given language variety, I have tried to collect as much information as possible, not only on presence or absence of the HON3pl-A, but also of honorific forms of direct address at the various locations. Unfortunately, since examples of direct address in dialect texts are even less common than of third person reference to elder relatives, the picture is far from complete. My general impression is that when speakers of a dialect use the HON3pl-A, they usually also use honorific forms in direct address, but not necessarily the other way round. 22 It is clear that in all Slavic languages in which the HON3pl-A is found, it is a thing of the past. Almost always, whenever there are testimonies of the use of the construction, it can be observed that it is used by the older generations and sometimes even rarely by them. The reverse is never true: there are no testimonies of the use of HON3pl-A as an innovation.
It is also clear that the extent to which the construction has died out in those areas where it probably existed varies between parts of the Slavic language area. For example, it is more alive in most of Slovak than in most of Croatian. Unfortunately we do not have ample material at our disposal for the whole Slavic language area for the last four centuries. The oldest testimonies date from the 1820s. If these testimonies reflect the speech of the oldest generations back then, we might be able to say something about the speakers born in those areas around 1750. However, most of the data given below come from dialect descriptions, atlases and text collections. The oldest informants whose speech is reflected in those sources were born around 1860, the great majority between 1880 and 1930. The dialect text collections do not show any correlation between the age of the informants interviewed at the various locations and the presence vs. absence of the HON3pl-A found there.
As a consequence of all that was said in the preceding two paragraphs • we cannot say exactly where the construction used to be found at an older stage of the different languages, and when and where it started to become obsolete; 21 Unfortunately for those interested in this specific phenomenon, many dialect anthologies (e.g. Vidoeski 2000 for Macedonian and Durnovo and Usakov 1910 for Russian) concentrate on tales, which, however valuable in other respects, rarely provide the contexts needed. In texts beginning with sentences like "There once was a father who had three sons", the father spoken about is not personally known to the speaker and therefore does not qualify for use of the HON3pl-A in the narrator's text. 22 For exceptions see Sect. 7.9 below.
• a negative example of use of the construction is a less strong signal than a positive one: if a speaker, even born in 1850, does not use the construction, it is very well possible that his / her mother or grandmother did.

Bosnian-Croatian-Serbian 23
In Burgenland Croatian the HON3pl-A is present in the northernmost dialects (spoken in Czechia and Slovakia). The remaining part of the Burgenland Croatian area (on both sides of the border between Austria and Hungary) shows a mixture of locations with and without the HON3pl-A.
In the Bosnian-Croatian-Serbian language variants spoken in the successor states of former Yugoslavia, the general picture shows an absence of the HON3pl-A.
I found two sources in which generalizations are made based on the presence of the HON3pl-A over a somewhat wider area: (a) Lončarić observes that most Kajkavian dialects have the HON3pl-A, though it is becoming obsolete (1996, p. 119). 24 Therefore this area is marked by a plus sign near Zagreb on Map 1. As can also be seen on this map, Lončarić's observation is corroborated by a couple of locations not far from this plus sign where positive or mixed examples were found (the Rs in the positive examples are generally elder relatives). The two triangles farther to the southwest on the border with Slovenia also represent observations on Kajkavian dialects (Lisac 2006, p. 115). No indication of the set of Rs is given here.
(b) Lisac remarks in his monograph on Čakavian that the HON3pl-A is used in Central Čakavian and gives an example (in which R is a mother) from Novalja on the island of Pag (2009, p. 113). However, I think that for two reasons Lisac's observation needs substantiation. First, he does not give any examples from other locations than Novalja. Second, I myself did not notice any use of the HON3pl-A in Central Čakavian during my fieldwork on Lošinj and Pag (including Novalja) and during my work on dialect material from Ugljan (see Budovskaja and Houtzagers 1994;Houtzagers and Budovskaja 1996), 25 and, as can be seen on Map 1, my findings from dialect texts and descriptions from other locations around Novalja are all negative. For that reason I chose to mark Novalja with a circle instead of a plus sign.
Farther to the south on the Dalmatian coast there is another positive mark on the map, which I put there with some hesitation due to the unclear formulation in the source, which also lacks a specification of the set of Rs (Geić 1996, p. 155). This mark represents the South Čakavian dialect of Trogir. 23 About the use and distribution of the HON3pl-A in Burgenland Croatian dialects there is an article by the present author (2012), to which I refer the reader for sources on the various sites. On the remaining part of Bosnian-Serbian-Croatian I consulted many volumes of Zbornik za narodni život i običaje (južnih slavena) (from 1896 to present), Barac-Grum (1993), Bogdanović (1979), Houtzagers (1985), Ivić (1964), Ivić et al. (1997), Jurišić (1966, Kalsbeek (1998), Lisac (2003, 2006), Lončarić (1996, Lukežić (1998), Lukežić and Zubčić (2007), Maričić Kukljičanin (2000), Nikolić (2000), Peco (1980), Sekereš (1976), Sławski (1962), Sobolev (1998), Težak (1981), Toma (1998), andVranić (2002). 24 Although Lončarić makes generalizations on 'most Kajkavian dialects', the contents of his observations, including his examples, are identical to Šojat's (1982, pp. 428-429) observations on the Turopolje dialects (with examples from Mraclin). The observations regard not only plural reference to a third person but also direct addressing with the masculine third person plural pronoun oni. There is no description of the set of possible Rs. The example sentences almost all have oni as their subject, three have gospon 'Sir' and gospa 'Madam', so that we could have to do with a HON3pl-C or a combination of HON3pl-A and -C. An early example of use of the HON3pl-A from the Varaždin region (in the northern corner of Croatia, north of Zagreb, exact location not given) can be found in Plohl-Herdvigov (1876, pp. 12-13). 25 There is of course a chance that I missed it.
In the remainder of the Bosnian-Croatian-Serbian language area only negative examples were found.

Slovene 26
Slovene is one of the two, perhaps three Slavic languages (the others are Slovak, perhaps Ukrainian) in which HON3pl-A is / has been present not only in the dialects, but also in the standard language. Nineteenth century grammarians like Murko (1843) advised 27 that one use the HON3pl-A when speaking about persons to which one wishes to show respect. In direct address the situation was more complicated: one should say vi (second person plural) to respected persons one is familiar with and oni (third person plural) to respected persons one is not familiar with (ibid., pp. 58-59). Toporišič (1974, p. 144) states that the third person plural referring to a single respected person who is absent is almost extinct. The examples he gives involve a father, a priest and a teacher. Of course one should have more examples in context, but it is possible that the situation that Toporišić describes is a combination of HON3pl-A and HON3pl-C. According to Reindl (2007), the construction is most often used when the R is a priest, grandparent or parent, "although in the past it could apply to members of the gentry as well" (ibid., p. 156). This also suggests a combination of HON3pl-A and HON3pl-C, the HON3pl-C part of which became obsolete first. With regard to the vitality of the construction, Reindl (ibid., p. 151) observes that it was used "until recently". 28 However, non-use of the HON3pl-A has been possible for a longer period, as we can see from the examples given by Miklosich (1883, p. 51). 29 In Slovene linguistic literature the term for honorific use of the third person plural (in direct address or referring to a third person) is onikanje (from oni 'they'; also Toporišič 1974, p. 144).
Let us now look at the dialects. The data points for Slovene on Map 1 are not optimally spread, but the map suggests the presence of the HON3pl-A across the whole Slovene language territory (along with absence in the centre), except for the Slovene-Italian border area, where an absence of the construction was found. There is one southern Slovene dialect with use of the HON3pl-A spoken on the Croatian side of the border. This is the circle with a dot in the middle on Map 1. 30 The plus sign in the middle of Slovenia marks the region of Carniola (Kranjska), where Metelko observed the use of HON3pl-A as early as 1825 (p. 224). The positive examples found in dialect texts and descriptions involve elder relatives and priests. Pronk explicitly defines elder relatives and the clergy as the class of Rs for the dialect of Egg and Potschach in Austria (2009, p. 149). At one location in the south of Slovenia (Zagorje na Pivki) negative examples for the HON3pl-A were found along with polite direct address to a non-relative with the third person plural pronoun oni. 26 On the use of the HON3pl-A in Slovene I consulted Lipavic Oštir (2011), Logar (1993), Metelko (1825), Miklosich (1883, Murko (1843), Pronk (2009), Reindl (2007, Sławski (1962), Steenwijk (1992), Toporišič (1974, and Zuljan Kumar (2007). 27 Murko presents his grammar as a description of the Volkssprache of all Slovenes (cf. the title of his work and his preface (no page numbers), but the fact that his description does not include any geographic variation suggests that he describes a variant of Slovene that he prefers. 28 According to Lipavic Oštir HON3pl-A, when used today, "seems archaic, or it can acquire ironic undertones" (2011, p. 43

Macedonian and Bulgarian 31
For the Macedonian and Bulgarian language areas (see Map 2) I found only negative examples. I also did not find any description in which the construction was mentioned for these languages.

Czech and Sorbian
The use of the HON3pl-A in Czech dialects is discussed by Bělič (1972) and Michálková (1971), with examples from different locations. Bělič observes that the construction can be found in the speech of the oldest speakers in Moravian and Silesian Czech and sporadically in east and southern Bohemia, although unfortunately, he does not provide any examples from 31 For Macedonian and Bulgarian I consulted Groen (1977), Hendriks (1976), Sławski (1962), Stojkov (1950), Vidoeski (2000 and two websites: Audio materijali od makedonskite dijalekti and Bulgarian dialectology as living tradition. According to Bělič, in some dialects attributive specifiers and adjectives and predicative adjectives can either agree in number and gender with the subject noun or be in the animate masculine plural, for example (9), but also (10) Michálková (1971) only reports on east Moravian dialects. According to her, the construction can also be used in direct address. What she says about agreement corresponds to Bělič's observations. She does specify the class of Rs: elder relatives and people that are referred to with a title or function. She states that the construction is chiefly found in the speech of the older generations, but that on the other hand one could occasionally hear children using it (1971, pp. 145-146). 32 By far the most positive and negative examples on Map 3 come from the text collection published by Michálková (1976). The picture given by Bělič (1972) and Michálková (1971)  Only the latter has positive examples of the HON3pl-A (the Rs are a mother and an aunt). The ancestors of the speakers of this dialect seem chiefly to originate from the southeast of Moravia, the emigrations took place in the middle of the 19th century (Pavlásek 2013, p. 152).
In Czech, direct polite address with oni 'they' is usually called oníkaní; referring to a single third person with a third person plural is called plurál úcty 'plural of respect'.
The people I consulted about Sorbian did not know the HON3pl-A at all. I did not check any texts myself.

Slovak
Slovak seems to be the language in which HON3pl-A is most alive. All of my informants on Slovak knew the construction, in most cases from rural settings and from contact with older speakers. One Slovak woman even heard a young doctor using it in a large Slovak town when referring to her old aunt: (11) Ale dobre vyzerajú. look.pres.3pl 'But she looks well.' This agrees with Ružička, who states that HON3pl-A is widely used in spoken Slovak, not only in rural communities but also in the towns (1957, pp. 86-87). 33 According to Orlovský, the HON3pl-A is used 'in some regions of Slovakia ' (1971, p. 123). This is confirmed by the data that I shall present below in this section in the sense that the presence of the HON3pl-A shows geographic variation, but the areas in which the HON3pl-A is present are spread all over the whole Slovak language territory. Regarding the class of Rs, Ružička primarily gives examples with elder relatives and later adds 'official persons and the intelligentsia in general' (1957, p. 87). Orlovský does not give a definition but gives examples with elder relatives, a school teacher and slečinka 'Miss (diminutive) ' (1971, p. 123). Mistrík adds that sometimes God is referred to using the HON3pl-A (see fn. 4): (12) Všetko nám pánbožko na úžitok stvorili.
create.l-part.masc.pl 'The good Lord (diminutive) created everything for our benefit.' (Mistrík 1981, p. 264) It is also interesting that the online Slovak dictionary slovník.sk lists both direct and indirect address of a single person by means of the third person plural under the lemma on 'he'. It is not labelled as dialectal. Direct address is characterized as obsolete, indirect address as obsolescent and belonging to the spoken language. On Map 4 the examples that I found in dialect texts are presented. Almost all of them are from the dialect subcorpus of the online Slovenský národný korpus, 34 in which at the moment some areas are overrepresented and others underrepresented. 35 This lack of balance can be seen from the map, but it is also obvious that positive examples are well-spread over the Slovak language area. The overwhelming majority of positive examples involved elder relatives. With respect to the other Rs potentially referred to using the HON3pl, there were some interesting incongruencies. Some dialects which show a nonuse of the HON3pl-A when referring to elder relatives do have positive examples involving persons with a high social status or Jesus Christ. 36 There are also examples for which the reverse is true, but that is less surprising because elder relatives and priests generally form the stable part of the set of Rs referred to by the HON3pl-A (see the end of Sect. 5 above). 33 We would now probably have to add 'by older generations of speakers', but it is possible that that restriction did not apply (or to a lesser extent) in 1975. 34 I also found a few examples in Vážný (1934).

Polish
The existence of HON3pl-A in Polish, especially in the south (the surroundings of Kraków and Rzeszów) has been documented by several linguists. 37 I refer the reader to Sikora (1993), Makarski (1973), Berger ( , 1996 and the sources used by those authors. Map 5 is based on examples that I found in Nitsch (1968) and Bąk (1974). The plus sign in the southeast represents the region around Rzeszów (see Makarski 1973). As can be seen from Map 5, the HON3pl-A can be found in the south of Poland, from Upper Silesia in the west to the Rzeszów area in the east and north of that area. 38 As we shall see in the in Sect. 7.7, the area with positive examples in the east continues into the Ukrainian language area. Again, the positive examples chiefly involve elder family members, sometimes priests. In one case a speaker refers to the deceased husband of her interlocutor using the HON3pl. In the same conversation HON3pl-A is also used for a priest, but not for a notary.
As is usual in works on Polish dialects, the west of Poland (including Middle and Lower Silesia in the south) is not covered. 39

Ukrainian
My information about the use of the HON3pl-A in Ukrainian is almost solely based on what I found in the Ukrainian dialect atlas by Nazarova et al. (1984Nazarova et al. ( -2001. 40 The atlas consists of three volumes that each cover a part of the Ukrainian-speaking area. In the second volume, which is about the westernmost third of the territory, one map (nr. 270, with comments on p. 59) is entirely dedicated to HON3pl-A. No such maps are given in the other volumes, but volume 3 offers some observations about the occurrence of HON3pl-A in the language area as a whole (p. 181, remark nr. 239). Volume 1 covers the central area, volume 3 the east and 37 An example from Upper Silesia can already be found in Roger (1863, p. 85). 38 Potebnja (1888, p. 5, footnote) seems to suggest that the HON3pl-A can be found in Slovincian (Baltic Sea coast). He refers to Miklosich (1883, p. 51), but there nothing can be found about Slovincian. 39 See for instance the maps in Nitsch (1968) and Bąk (1974). This is because the Polish speaking population in those areas moved there relatively recently. 40 With the exception of one example that I found in Mel'ničenko (1985, p. 149). Map 270 of Nazarova et al. (1984 2) is covered with a multitude of data points with positive, negative and mixed examples. Because of the density of the data points on the map I cannot reproduce it on Map 6. Instead I chose to place a plus sign in the middle of that subarea. This is because positive and mixed examples form the great majority. What is not shown on Map 6 is that toward the south and especially the north the number of negative examples gradually increases. In approximately the northernmost quarter of this subarea (a part of that quarter lies north of the Ukrainian-Belorusian border) negative examples form a majority. This is continued by Belarusian, which, as we shall see in Sect. 7.8, shows only negative examples. In a number of locations fieldworkers remarked that the form was used by (middle-aged and) older speakers, in one case the reverse was observed, which could be accidental (Nazarova et al. 1984). An example of the use of the HON3pl-A in the Char'kov oblast can be found as early as 1857 (Kuliš 1857, p. 283).
About the central area, covered by volume 1 of the atlas, the authors state that the HON3pl-A is sporadically heard in the speech of the older generations and a couple of individual locations are given where the HON3pl-A was attested (Nazarova et al. 1984. These locations are visible on Map 6: they stretch to the very east of this subarea. I find it hard to believe that this is a realistic picture of the situation at the time when the data for the atlas were gathered. The map of the western area shows no decrease of presence of the HON3pl-A toward the eastern border, which is almost a straight vertical line. I expect that if a map showing the presence of the HON3pl-A had been made for the central area as well, we would have seen a gradual decline of the use of the construction, probably related to the age of the speakers. Due to the fact that the HON3pl-A is sporadically heard in the speech of the older generation, and because we are looking for traces of a disappearing phenomenon, I regard this central subarea as a whole as 'positive', albeit not as positive as the western one. Therefore I marked this subarea with a plus sign as well. The authors of the atlas state that an absence of the HON3pl-A is typical for the area covered by the third volume, except for two locations where they attested it (Nazarova et al. 1984. I marked this subarea with two minuses (east and south). We do not get a clear picture of the set of potential Rs from the atlas. The editors of the atlas write that they want to make an inventory of the use of the množyna povagy (plural of respect). In their examples the Rs are restricted to different words for 'mother' and 'father'. Shevelov (1963, p. 67) states that HON3pl-A is used "even now in some peasant dialects; in literary language it is regarded as archaic" and he gives two examples from 19th century authors, in which the Rs are a father and "master L'ol'o from the neighboring farm". This suggests that the construction was once, at least to some degree, accepted in literary language.
Map 6 contains two Ukrainian dialects spoken in the Vojvodina, one of which has the HON3pl-A. This site is of potential interest for establishing the age of the phenomenon. Karskij (1956, p. 332) describes the use of the HON3pl-A in Belarusian, suggesting that his account is valid for the language as a whole. He observes that sometimes children refer to respected persons in their absence using plural forms. In his examples the Rs are always a parent of the speaker. He mentions two locations, one of which is Białystok in present-day Poland. 41 Contrary to his account, I have not found a single positive example in the dialect text anthology edited by Padlužny (1990). Therefore all marks on Map 7 are negative, except for Białystok in the west.

Russian 42
In Sect. 5 I have argued that Russian has / had a honorific construction that I called 'HON3pl-C' and that would be worth further study. 41 I have not been able to trace the other location, which is abbreviated 'Rast.' 42 For Russian I consulted, apart from the publications mentioned in this section, Durnovo andUsakov (1910), Gerd (2002), Kasatkina (1991), Mel'ničenko (1985, Trubinskij (2004), Xonselaar (2001, the online corpus Nacional'nyj korpus russkogo jazyka and a collection of unpublished South Russian dialect texts sent to me by courtesy of Leonid Kasatkin. Map 7 Belarusian. One triangle (mixed), for the rest only black dots (negative) As can be seen on Map 8, my Russian findings for the HON3pl-A are almost exclusively negative. However, there are a few data that deserve our attention. 43 Petrovskaja (1922;apud Šapiro 1953, p. 181) gives an example from Matyra (Moskovskaja oblast') in which, oddly enough, the singular pronoun on 'he' is used with a plural verbal predicate: (13) Vot on prizyvajut' mamu. call.pres.3pl 'There he calls mother.' Although Šapiro does not reveal who the R is in this example, he does point out that it is a person who deserves special respect and that there are more examples from Southern Russian dialects (1953, p. 181).
Possibly even more interesting are Steinke's texts from the Russian speech island Kazaško in Bulgaria near Varna. Text nr. 2, short as it is, contains many examples of the HON3pl-A, in which the Rs are the father and mother of the speaker (1990, p. 234). According to Steinke, the dialect was originally Southern Russian and was separated from the Russian language area between 1750 and 1800 (1990, pp. 80-81). This is not only one more indication that the HON3pl-A was present in (parts of) Southern Russian but also seems to provide us with a terminus ante quem for HON3pl-A in (some parts of) Southern Russian.
And there is more. Berger (1996, p. 24) cites a fragment from a dialect text from Karelia (the exact location is not given) with a clear example of HON3pl-A referring to a father (along with a couple of negative examples). I found two more sites in the far north, one in Karelia and one in the Murmansk oblast some 16 kilometers from the Karelian border, with positive examples for a priest, a grandfather and a father (along with negative examples from both sites; see Gerd (2002, pp. 172, 174, 189, 194).
Most of the Russian-speaking sites where examples for the HON3pl-A were found show atypical characteristics: Matyra has a singular personal pronoun in the subject position with a  (13), Kazaško has no distinction between formal and informal direct address (Steinke 1990, p. 157 fn. 48) and the northernmost location with HON3pl-A shows instances of direct address to respected persons by means of the second person singular.
All this seems to indicate that it is desirable to find out more about the Russian language area, especially its south and far north. Since it is not to be expected (though not excluded) that in these areas even the oldest speakers alive will produce HON3pl-A forms, it is to be hoped that more data can be found in lesser-known or unpublished sources.
In the substandard speech variant of the city of Perm' (see Skitova 1989) and in dialects east of Perm' (see Corbett 2009, p. 346) there exists a construction that is syntactically similar to HON3pl-A but refers to a person (not necessarily a respected person) and one or more others: 44 (14) Tanja ešče pridut. (Skitova 1989, p. 131) come (perfective).pres.3pl The examples found for non-European Russia in Mel'ničenko (1985) are all negative and the sites where they were found stretch out very far eastward. For obvious reasons I did not include that part of Russia in Map 8. I inserted arrows to point to the two locations in the north, which could easily be missed by the reader.

Overview
Map 9 is meant to give the reader an impression of the distribution of the HON3pl-A in Slavic as a whole. Because of the large amount of data to be shown on a relatively small map I omitted the negative data points and chose black dots to represent the positive and mixed ones. Some remote locations that could escape the reader's notice are indicated using arrows.

Conclusions
• In a number of Slavic languages there exists or has existed a possibility of referring to a single person not present to the conversation and respected by the speaker by way of a construction that I have called 'HON3pl-A'. If the respected person is the subject of the sentence, the predicate is in the plural. If the respected person is referred to with a pronoun, this pronoun has a plural form. • HON3pl-A is almost exclusively found in dialects, only in Slovene, Slovak and perhaps Ukrainian has it to a certain degree penetrated into literary language.
• The class of persons referred to by HON3pl-A can be described as: 'persons who are respected by the speaker in the affectionate way in which a child respects (or is by tradition bound to respect) an elder relative'. In the vast majority of attested cases the HON3pl-A refers to elder relatives or priests. Sometimes other respected persons are referred to (e.g. persons who have a higher social position), but this subclass of referents seem to be less stable. • HON3pl-A has typologically interesting characteristics: great horizontal distance is contraindicative and warmth and closeness play an important role. • There are also other cases of reference to a single person using a third person plural: (a) persons not participating in, but present while the conversation is taking place or (b) persons with a high social rank not present while the conversation is taking place. I called these 'HON3pl-B' and 'HON3pl-C', respectively. HON3pl-A, HON3pl-B and HON3pl-C sometimes coexist in the same language varieties, but must be seen as separate phenomena. • HON3pl-A is / was found in (a) a more or less continuous area in which Slovene, Kajkavian Croatian, Burgenland Croatian, Moravian and Silesian Czech, Slovak, southern and east Polish, western and central Ukrainian are spoken. There are some indications that it existed in Southern Russian as well. I call this area 'more or less continuous' because it also contains locations (spread all over the area) from which we have only negative examples; (b) some scattered locations in Croatia and the middle and north of European Russia; (c) three speech islands: one on the Bulgarian coast (southern Russian), one in the west of Romania (Czech) and one in the Vojvodina (Ukrainian). It is not found in Sorbian, Bohemian Czech, most of Polish, most of Bosnian-Croatian-Serbian, Macedonian, Bulgarian, eastern and southern Ukrainian, Belarusian and almost all of Russian.

Thoughts about future research
All questions relating to the origin of the construction remain to be answered. Where and when did the HON3pl-A arise in Slavic? Did it arise under the influence of another language? If it first arose only in part of the 'more or less continuous area' described in Sect. 8, how, when and why did it spread over the remainder of that area? Head (1978, p. 163) considers variation in number as a means of showing respect to a third person a universal tendency and argues that there is no need to account for this phenomenon by special circumstances in the languages or language communities in which it occurs. Yet I think that the relative compactness of the area where the HON3pl-A can be found and the specific characteristics of the HON3pl-A that are shared all over that area justify these questions. Finding answers will not be simple. One of the causes for this is the lack of dialect data for the period that is presumably relevant. It is improbable that the HON3pl-A is a Common Slavic phenomenon. If it were, one would expect to find at least traces of it in the oldest documents, such as the Novgorod birchbark letters. On the other hand we can assume (not without some degree of speculation) that it was present, at least in some places, around 1750 (see Sect 7.9 above). In order to corroborate this and to find out more about the rise of the construction we would need dialect material from 1750 and well before, which is difficult to find.
If we try to explain the rise of the HON3pl-A having been a result of influence from other languages, there are also some obstacles to overcome. The contact language one thinks of first is, of course, German. German could have influenced Slavic directly by its HON3pl-C (of which we do not seem to know the chronology, see Sect. 5 above), indirectly by its polite direct address by means of the third person plural (siezen), or indirectly at an earlier stage by its having a distinction between informal du (second person singular) and formal ihr (second person plural). There are many conceivable scenarios. But whatever scenario we choose, we shall have to explain among other things why in some areas with intense exposure to German the HON3pl-A is not found (such as Sorbian, Bohemian Czech 45 and parts of Polish), while in other areas with little or no exposure to German (such as central Ukrainian) it is.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.