stance, cannot be derived from good, because phonetic laws will not allow it, and because the two words run parallel, and never approach one another, as far as we can follow their history.

But even where an etymology is unassailable on phonetic and historical grounds, it can never give us more than the first starting-point of a word. It may teach us how the object to be named was first conceived, but no more. We know, for instance, that deus in Latin represents the Sanskrit deva, perhaps also the Greek Oeós, though neither of these etymologies is in strict accordance with phonetic rules, and that deva meant originally, bright. This is extremely important as showing us that one of the many conceptions of the Divine started from the concept of bright and beneficent beings, such as sun, and moon, and stars, in opposition to the dark and deadly aspects of the night; but to imagine that this could help us to understand the concept of God in the mind of such a thinker as Pascal, would be absurd. can never be too grateful, if we can discover the germinal idea of a word, if we can prove, for instance, that deus was originally no more than a bright being, that a priest was originally an elder, a minister a servant, a bishop an overseer; but if we were to give these etymologies as more than historical curiosities, and mistake them for definitions, we should only prove our ignorance of the nature of language, which is in a constant state of ebb and flow, and exhibits to us the process of continuous evolution better than any other part of nature.

We

1 See Selected Essays, i. p. 215. I still hold to the opinions there expressed.

Historical Definition,

We now come to historical definitions. What I call an historical definition is an account of these very changes which take place in the meaning of a word, so long as it is left to the silent and unconscious influences which proceed from the vast community of the speakers of one and the same language. Thus an historical definition of deus would have to show the various changes which led from deva, bright, as applied to the sun, the dawn and other heavenly phenomena, to the Devas, as powers within or behind these heavenly bodies, and lastly to the beneficent agents in nature or above nature, whom the Hindus called Devas, and the Romans dii. As the biography of a man may be called his best definition, what I call biographies of words are perhaps the most useful definitions which it is in our power to give.

Dogmatic Definition.

Lastly come the dogmatic definitions, by which I mean definitions given on the authority of individuals, who, whatever a word may have meant etymologically, and whatever it may have come to mean historically, declare that, for their own purposes, they intend to use it in such and such a sense. This is chiefly done by philosophers, lawyers, and men of science, who feel unable to use important words with all the vagueness of their etymological and historical meaning, and determine once for all, generally by the old logical method of settling their genus and their specific difference, in what exact sense they ought to be employed in future.

Let us now see how these three kinds of definition

have been applied to the word with which we have to deal, namely religion.

Etymological Definition of Religio.

The etymological definition of religion has attracted considerable interest among theologians, owing to that kind of tacit persuasion that the etymology of a word must somehow or other help to disclose its real meaning. It is well known that Lactantius derived religio from religare, to bind or hold back, and he did so, not simply as a philologist, but as a theologian. 'We are born,' he says, 'under the condition that, when born, we should offer to God our justly due services, should know Him only, and follow Him only. We are tied to God and bound to Him (religati) by the bond of piety, and from this has religion itself received its name, and not, as Cicero has interpreted it, from attention (a relegendo)1.'

Before we examine this etymology, it will be useful to give the etymology which Lactantius ascribes to Cicero, and which he is bold enough to reject. Cicero says: Those who carefully took in hand all things pertaining to the worship of the gods, were called religiosi, from relegere, -as neat people (elegantes) were so called from elegere, to pick out; likewise diligent people, diligentes, from diligere, to choose, to value, and intelligent people from intel

1 Lactantius, Institut. Div. iv. 28, 'Hac conditione gignimur, ut generati nos Deo justa et debita obsequia praebeamus, hunc solum noverimus, hunc sequamur. Hoc vinculo pietatis obstricti Deo et religati sumus; unde ipsa religio nomen accepit, non, ut Cicero interpretatus est, a relegendo.'

Rather from a lost verb elegare.

D

ligere, to understand; for in all these words there is the meaning of legere, to gather, to choose, the same as in religiosus1.

Let us first clear the ground of some statements which are repeated again and again, but which have really no foundation. It is often said that Varro 2 supports the etymology of Lactantius, but Varro simply treats of legere and legio, and thus supports indirectly the etymology of Cicero, rather than that of Lactantius.

Festus, again, if he is to be quoted at all as having given an etymology of religio, sides with Cicero, and not with Lactantius, for he says that people are called religiosi if they make a choice (delectus) of what has to be done or to be omitted in the worship of the gods, according to the custom of the state, and do not entangle themselves in superstitions 3.

Of later writers St. Augustin follows sometimes the one, sometimes the other derivation, as it suits his purpose; while among modern theologians it has actually been maintained that religio was descended from religare as well as from relegere, so as to combine the meanings of both *.

From a purely philological point of view it cannot

1 Cicero, De Nat. Deor. ii. 28, 'Qui autem omnia quae ad cultum deorum pertinerent diligenter retractarent et tamquam relegerent sunt dicti religiosi ex relegendo, ut elegantes ex eligendo, itemque ex diligendo diligentes, et intelligendo intelligentes. His enim in verbis omnibus inest vis legendi eadem quae in religioso.' "Varro, De ling. lat. v. 68; ed. Egger. Legio, quod leguntur milites in delectu;' Nitzsch, Studien und Kritiken, i. p. 527.

Festus, p. 236, 'Religiosi dicuntur, qui faciendarum praetermittendarumque rerum divinarum secundum morem civitatis delectum habent nec se superstitionibus implicant.'

1. c.

'Relegendo se sentit religatum,' von Drey, as quoted by Nitzsch,

be denied that religio might have sprung from religare quite as well as from relegere. The ordinary objection that from religare we should have religatio, and not religio, has no real weight, for we find by the side of opinari such words as opinio, not opinatio, and necopinus; and by the side of rebellare, rebellis and rebellio. In lictor also, if it meant originally a man who binds the criminal, we should have to admit a root ligere, by the side of ligare.

The real objection to our deriving religio from religare is the fact that in classical Latin religare is never used in the sense of binding or holding back. In that sense we should have expected obligatio, or possibly obligio, but not religio. Cicero's etymology is therefore decidedly preferable, as more in accordance with Latin idiom. Relegere would be the opposite of neglegere or negligere1, and as neglegere meant 'not to care,' relegere would naturally have meant 'to care,' 'to regard,'' to revere'. From a verse quoted by Nigidius Figulus from an ancient writer, and preserved by Gellius (iv. 9), we learn that religens was actually used, as opposed to religiosus. He said: Religentem esse oportet, religiosus ne fuas, 'it is right to be reverent, but do not be religious,' that is, superstitious 3.

1 The change of e into i is historical. We find neglego and negligo, intellego and intelligo. The spelling with e is the old spelling, but there are modern compounds also which have always e, such as perlego, praeligo.

2 M. M. Hibbert Lectures, p. 22.

3 Gellius, ed. Hertz, iv. 9. Adjectives in osus generally imply an excess, as vinosus, mulierosus. Thus Nigidius Figulus said: 'Hoc inclinamentum semper hujuscemodi verborum, ut vinosus, mulierosus, religiosus significat copiam quandam immodicam rei super qua dicitur. Quocirca religiosus is appellabatur qui nimia et superstitiosa religione sese alligaverat, eaque res vitio assignabatur.' 'Sed