It is widely believed (and with good reason) that reciprocal pronouns such as English each other require plural antecedents:

(1)	a.		The children envy each other.
	b.		Sally and Anna envy each other.
	c.	*	Joe envies each other.

It is a central goal of the analysis of reciprocals presented in Heim, Lasnik, and May (1991) to provide an understanding of why this requirement should hold. The analysis they develop takes the reciprocal to be not an atomic lexical element but rather to be composed of its two apparent parts—each and other. The element each is identified in their proposal with the distributive quantifier each as it appears in (2):

(2)	a.		The children have each bought a book.
	b.	*	The class has each bought a book.

The requirement of plurality in (1) then reduces to the similar requirement imposed by distributive each in (2), whose semantics also plays a crucial role in the semantics of reciprocals that Heim, Lasnik, and May offer.

The Irish reciprocal is the element a chéile. In origin, this is a complex expression consisting of a possessive pronoun a ('his' or 'its') followed by the noun céile, which means 'partner,' 'mate' or 'spouse'.

(3)	Chonaic	siad	a chéile.
	saw	they	each other
	'They saw each other.'

Although this decomposition is diachronically valid, it seems that in the modern language a chéile is unanalyzable. The particle a does not vary according to attributes of the antecedent (see (3) in which a retains its apparently singular form even in the presence of a plural antecedent); furthermore, the long [e:] characteristic of the noun céile is reduced and neutralized in the reciprocal. There is, in any case, no (visible) sign of a marker of distributivity within the reciprocal. If one were pushed to provide a literal translation of the expression, it would be something like 'the other'.

The restrictions which limit the distribution of a chéile in Irish closely mirror those which govern the distribution of English reciprocals. I will not document this in detail here, but simply assert that the antecedent must be more prominent than the reciprocal and must be in a local relation with it. The issues that arise in cashing out such general statements are closely parallel, down to minute levels of detail, in Irish and in English. Some of the core patterns are illustrated very briefly in (4) and (5)¹.

(4)	a.		Labhair	siad	le-n	a chéile.
			spoke	they	with	each other
			'They spoke to each other.'

	b.		Cháin	siad	leabhair	a chéile.
			criticized	they	books	each other
			'They criticized each other's books.'

	c.		Shuigh	siad	in aice le-n	a chéile.
			sat	they	beside	each other
			'They sat beside each other.'

(5)	a.	*	Dúirt	siad	go	raibh	a chéile.	breoite.
			said	they	COMP	was	each other	ill
			'They said that each other were ill.'

	b.	*	Dúirt	siad	go	bhfaca	Eoghan	a chéile.
			said	they	COMP	saw	Owen	each other
			'They said that Owen saw each other.'

So far the picture is a familiar one. Unexpectedly, however, the Irish reciprocal permits singular antecedents, at least in a range of circumstances. Examples on the model of (1c) are, of course, as ungrammatical in Irish as they are in English:

(1c')	*	Chonaic	Eoghan	a chéile.
		saw	Owen	each other
		'Owen saw each other.'

Consider, though, the examples in (6)–(12) (all attested examples):

(6)	D'fhill	Sáile	a	clóca	ar	a chéile
	folded	Sally	her	cloak	on	each other
	'Sally folded her cloak up (lit.: on each other).'

(7)	Dhún	sé	an	páipéar	ar	a chéile
	closed	he	the	paper	on	each other
	'He closed the paper up (lit.on each other).'

(8)	D'fhill	sé	an	nóta	ar	a chéile
	folded	he	the	note	on	each other
	'He folded the note up (lit. on each other).'

(9)	Chas	sí	an	léine	ar	a chéile
	turned	she	the	shirt	on	each other
	'She folded the shirt.'

(10)	D'fháisg	sí	a	béal	ar	a chéile
	squeezed	she	her	mouth	on	each other
	'She squeezed her lips together.'

(11)	Bhí	a	cár	ag greadadh	ar	a chéile
	was	her	teeth (coll.sg.)	beating	on	each other
	'Her teeth were chattering.'

(12)	Má	choinním	mo	bhéal	ar	a chéile
	if	I-keep	my	mouth	on	each other
	'If I keep my mouth shut.'

The interesting property of (6)–(12) is that even though they all involve singular antecedents, they are still reciprocal in their interpretation (the translations in terms of 'each other' are not entirely fanciful). The reciprocal relations involved though, are not reciprocal relations which hold among individuals conceived of as indivisible wholes. They are rather reciprocal relations which hold among parts of an individual whole. The folding of a cloak, of a piece of paper, or of a shirt (as in (6)–(9)) involves conceiving of the cloak, shirt, or sheet of paper as a set of parts, each of which is brought into contact with the other. Similarly in (10)–(12), the antecedent noun is singular (béal meaning 'mouth', cár meaning 'set of teeth' or 'jaw'), but the interpretation again crucially involves reciprocal relations among parts of a whole.

It becomes clearer that this is the right interpretation of the semantics when one notices examples like (13)–(15):²

(13)	na	gadhair	ag streachailt	an	ghiorria	bocht	ó	chéile
	the	hounds	tearing	the	hare	poor	from	each other
	'the hounds tearing the poor hare apart'

(14)	scar	sé	ó	chéile	an	páipéar	rua
	separated	he	from	each other	the	paper	red
	'He tore the red paper apart.'

(15)	go	raibh	sí	scoite	ó	chéile
	COMP	was	it	disconnected	from	each other
	'that it had been broken apart'

In these cases, the English translation which most naturally renders the Irish reciprocal is the lexical item apart, an element which (on this use at least) requires reference to an individual conceived of as a sum of its parts, and which further requires that those parts be in motion outwards and away from one another. If something breaks, falls, comes, or blows apart, then each part of that thing must be in motion away from each other part.

The preposition ó meaning 'from' denotes motion away from the point of view. When it governs the reciprocal, then, it will denote reciprocal motion of parts outward from some central point of view—a very close approximation to the meaning of English apart. Naturally enough, different prepositions, when they govern the reciprocal, will yield interpretations which cannot be rendered by way of English apart, but which still entail reciprocal relations among parts of the whole. This possibility is illustrated for the preposition ar, corresponding very roughly to English 'on', in (16)–(18):

(16)	Tá	an	teach	titithe	anuas	ar	a chéile
	is	the	house	fallen	down	on	each other
	'The house has fallen in on itself.'

(17)	bhí	sé	ina chodladh	casta	ar	a chéile
	was	he	asleep	turned	on	each other
	'He was asleep, curled up (on himself).'

(18)	gach	putóg	im	bolg	casta	ar	a chéile
	every	gut	in-my	belly	twisted	on	each other
	'every gut in my belly twisted up'

Or with the preposition i(n) (meaning 'in') in (19):

(19)	Tiocfaidh	an	cnámh	i-n	a chéile
	will-come	the	bone	in	each other
	'The bone will knit.'

All of these examples have natural interpretations involving reciprocal relations among parts of a whole. In (16), each part of the house has fallen in towards each other part; in (17) the sleeper has each part of his body turned in towards each other part; in (18) each part of the gut is twisted towards each other part, and in (19) each part of the bone joins each other part.

What are we to make of all of this?

These observations suggest in the first place that there is no necessary or general requirement that reciprocal anaphors take only plural antecedents. Reciprocal anaphors may have singular antecedents in at least one circumstance—when that singular antecedent can be interpreted as a sum of parts and reciprocity can be asserted to hold of relations holding among those parts.

These observations may also suggest that Heim, Lasnik, and May are right in attributing the requirement that English reciprocals have plural antecedents to the presence of the distributive quantifier each within the (English) reciprocal phrase. In the absence of such an element (as in Irish it seems), the requirement does not always hold.

References

Heim, Irene, Howard Lasnik, and Robert May. 1991. 'Reciprocity and plurality.' Linguistic Inquiry 22: 63–101.

Endnotes

² In (13)–(15), the (prosodically weak) initial a of a chéile is elided following the long vowel of the preposition. [Back]