Reflexive pronouns (or “reflexive anaphors“) are expressions which are prototypically used to indicate that a non-subject argument of a transitive predicate is coreferential with (or bound by) the subject, i.e. expressions like German sich, Russian sebja, Turkish kendi, Mandarin zìjĭ, English x-self:
John i saw himself i in the mirror.
By intensifiers we mean expressions like German selbst, Russian sam, Turkish kendi, Mandarin zìjĭ, English x-self, which can be adjoined to either NPs or VPs, are invariably focused and thus are prosodically prominent. The main function of intensifiers can be seen in the evoking of alternatives to the referent of the NP they relate to:
The director himself opened the letter.
The director opened the letter himself.
Our main motivation for considering reflexives and intensifiers in tandem is that they are frequently identical in form and thus only differentiated in terms of distribution (as, for instance, in English). In languages in which they are formally differentiated, intensifiers can be used to reinforce reflexive pronouns. This is illustrated by (3c) from German, where the reflexive pronoun (sich, cf. (3a)) and the intensifier (selbst, cf. (3b)) are formally distinct.
in the mirror
‘Johann saw himself in the mirror.’
‘The director himself came to welcome us.’
‘Paul criticized himsélf.’
The map shows the geographical distribution of the two types of languages:
|Go to map|
|Intensifiers and reflexive pronouns are formally identical||94|
|Intensifiers and reflexive pronouns are formally differentiated||74|
The correct identification of intensifier-reflexive identity encounters the following three problems. First, there may be only partial identity. A typical situation, exemplified by Malayalam (Dravidian; southern India) in (4), is that the reflexive expression is a combination of intensifier and (personal or possessive) pronoun. For the purposes of the map, cases of partial and complete identity have been lumped together and contrasted with cases of non-identity.
‘He accused himself.’
‘I myself will do it.’
Secondly, languages may have more than one reflexive pronoun, only one of which may be (partially or completely) identical to the intensifier. A case in point is Dutch, where different verbs require different reflexive pronouns:
Jan wast zich.
Jan haat zichzelf.
‘John hates himself.’
Thirdly, there are languages that have different intensifiers for adnominal and for adverbial use, and that only use one of these (typically the adverbial one) as reflexive. This is illustrated by Japanese (cf. 6a-c; Ogawa 1998: 165-173).
‘Taro washed the car himself.’
‘Taro himself honors the professor.’
‘Taro criticized himself.’
In the entire sample there are 61 languages that manifest complete identity of reflexive and intensifier. A further 33 languages show partial identity whereas 74 languages draw a clear formal distinction between the two expressions. The combination of cases of partial and complete identity yields 94 languages with intensifier-reflexive identity, contrasting with 74 languages that do not exhibit identity in this domain (56% versus 44%). Although there is thus a slight preference for identity, this parameter divides our sample into two parts of almost equal size.
Language families manifesting mainly intensifier-reflexive identity are Sino-Tibetan, Dravidian, Afro-Asiatic, Altaic, Trans-New Guinea and the Indic branch of the Indo-European family. Differentiation predominates in Uto-Aztecan, Penutian and in the European branches of Indo-European. Among the language families that have members illustrating both identity and differentiation we find Austronesian, Australian and Niger-Congo.
In terms of areal distribution there is a major divide, with Central Asia, East and Southeast Asia, the South Asian subcontinent and New Guinea showing a greater than average share of languages with identity, whereas Europe and the Americas mostly harbour languages that keep intensifiers and reflexives apart. Of course, the areal clustering in Europe can be attributed to genetic factors, but we seem to be dealing with a true Sprachbund phenomenon in Asia and in New Guinea. Identity is also found on the fringes of Europe, northern Africa and the Middle East. Mesoamerica, central Africa and Australia show no clear areal patterning.
Another parameter that we have tried to investigate concerns the lexical sources of intensifiers and reflexives. However due to a multitude of adverse factors (the lexical source is frequently not transparent, etymological dictionaries exist for only a few languages, grammar books only rarely give this kind of information) we have only been able to find this information for slightly over sixty languages. Therefore no separate map is given. The languages for which we do have reliable information show that intensifiers and reflexives typically derive from expressions for body parts (also including ‘soul’). Whether or not intensifiers are invariably an intermediate step in the development of reflexives from expressions for body parts is not clear in all cases.
Of the 62 languages for which we have information about the lexical source of the reflexive, 47 have expressions that are related to a body part noun:
2 languages (e.g. Modern Standard Arabic)
1 language (Modern Hebrew)
1 language (Dongolese Nubian)
1 language (Ngiti)
A high proportion of these are African languages. Other lexical sources include anti-comitative expressions like ‘alone’ (Lealao Chinantec), expressions indicating an extreme or a precise value ‘very, exact’ (Chacaltongo Mixtec), ‘return’ (Paamese) and ‘reflection in water’ (Finnish).
Of the 62 languages with a known lexical source, 33 have reflexive-intensifier identity, and in 24 of these 33 languages the intensifier/reflexive is related to a body part expression. Examples of languages whose reflexive pronoun, though distinct from the intensifier, is related to a body part expression include Basque (buru- ‘head’), and Ngiti (ndɨ̀ ‘skin’). Conversely, there are languages that have an intensifier related to a body part expression, but whose reflexive is derived from a different source (Rama: intensifier páin ‘body’, reflexive áp ‘alone’). There are two closely related languages in the sample that draw a formal distinction between reflexives and intensifiers, such that both of them derive from expressions for body parts, albeit different ones: Koyraboro Senni intensifier hundey ‘soul’ vs. reflexive boŋ ‘head’, Koyra Chiini intensifier huneyno ‘body’ vs. reflexive bomo ‘head’.
Distinguishing two types of languages on the basis of the relationship between intensifiers and reflexives is interesting and relevant in that identity vs. differentiation of intensifier and reflexive can be shown to correlate with other properties of reflexive pronouns. If a language uses the same expression both as intensifier and as reflexive pronoun, this expression is not used as a middle marker or marker of derived intransitivity (cf. König and Siemund 2000a: 59). Languages belonging to the same type as English never use reflexive markers in constructions like (7a-c) from German or (8a-b) from Spanish:
‘Charles sat down.’
‘The door opened.
‘This shirt washes well.’
‘Used cars for sale.’
‘Spanish spoken here.’
Other implicational connections that have been proposed and discussed in the literature relate to the possible antecedents of reflexives and to their binding domain. Thus reflexives that allow long-distance binding seem invariably to be different from intensifiers (Norwegian seg; Latin se, sibi), whereas reflexives of the English type can only be bound in a local domain. On the other hand, expressions that can only be used as reflexives seem to require a subject as antecedent, while expressions like English x-self also allow objects as antecedents. Such implicational connections have not been shown to hold, however, for a sufficiently large sample of languages.
Body part expressions can develop directly into reflexive pronouns as a result of metonymic change (cf. Faltz 1985). Intensifiers also play an important role in the genesis and renewal of reflexive markers. In example (9) from Old English (Norman 1848), the addition of the intensifier to the referentially independent object pronoun makes this pronoun co-referential with the subject, leading to the development of a complete paradigm of reflexives (similarly to Malayalam).
(9) Old English
‘Judas hanged himself.’
Another way of renewing reflexive markers is by intensifying weak reflexives. Such developments are widely attested in Romance languages, where the reflexive clitics are being replaced by the respective tonic forms in combination with intensifiers (cf. (10) from Portuguese; similarly in Swedish, Norwegian, Dutch).
‘Maria looks at herself in the mirror.’
‘He cut himself.’