No.Lancashire.Rosehearty.Notts.Morpeth.Gainford.Norfolk.Beddgelert.
1.Sally Waters.—Sallie [ ].——Sallie [ ].Sallie [ ].
2.—Sally Walker.—Sally Walker.Sally Walker.——
3.———————
4.———————
5.———————
6.———————
7.———————
8.———————
9.———————
10.——Sitting on the ground.———Sitting in sand.
11.———————
12.———————
13.Rise for a young man.Rise for a good man.——Rise and choose your good man.Rise and choose.—
14.——Sobbing for a young man.Lamenting for a young man.——Crying for a young man.
15.———————
16.———————
17.———————
18.First east, west.Choose east, west.Turn east, west.Choose east, west.Choose east, west.Choose east, west.—
19.———————
20.———————
21.Then to the bestloved.—Turn to the best loved.Choose the best loved.Choose the best loved.Choose the prettiest.Pick the one you like best.
22.Now you’re married, &c.There’s a couple, &c.—Here’s a couple, &c.Now they’re married, &c.Now you’re married, &c.Now you’re married, &c.
23.———————
24.———————
25.———————
26.———————
27.——A bogie in, &c.————

The first thing to note from this analysis are the words Sally and Water. In twenty-three versions they are Sally Water or Waters, in seventeen versions it is Sally Walker, in six versions it is another name altogether, while in two versions it is Sallie only. The most constant name, therefore, points to Sally Water as the oldest version; and it is noticeable that in the [Lincolnshire] and [Sheffield versions], where the name is not Sally Water, the word water is introduced later on in the line which directs the action of sprinkling water. Is it possible, then, that Sally Water may be a corruption from an earlier form where Sally is some other word, not the name of a girl, as it is usually supposed to be, and the word water is connected, not with the name of the maiden, but with the action of sprinkling which she is called upon to perform? If we could surmise that the early form was “Sallie, Sallie, water sprinkle in the pan,” the accusative being placed before the verb, the problem would be solved in this manner; but there is no warrant for this poetical licence in popular verses, and I prefer to suggest that “water” got attached as a surname by simple transposition, such as the [Norfolk] and [Beddgelert versions] allow as evidence. It follows from this that Walker and other names appear as degraded forms of the original, and do not enter into the question of origins, a point which may readily be conceded, considering that the general evidence of all these singing games is, that no special names are ever used, but that names change to suit the players. The next incident in the analysis is the ceremony of “sprinkling the water,” which is constant in twenty-one versions, while the [Wakefield] “Springin’ in the pan,” the [Settle] “Tinkle in a can,” [Halliwell’s] “Sprinkle for a young man,” and the eight versions in which this incident is wholly absent in any form, are evident corruptions. The tendency of the corruption is shown by this to be that the “sprinkling of water” came to be omitted from the verse, and therefore the other variants—

are but the steps through which the entire omission of the water incident was finally attained. The third incident is “Rise and choose” a young man, the alternative being “Crying for a young man.” The first indicates a kneeling and reverential attitude before the water, and occurs in twenty-one versions, while the second only occurs in fourteen versions.

The expression “crying” is really to “announce a want,” as “wants” were formerly cried by the official “crier” of every township, and indeed as children still in games “cry” the forfeits; but losing this meaning, the expression came to mean crying in the sense of “weeping,” and appearing to the minds of children as a natural way of expressing a want, would therefore succeed in ousting any more archaic notion. The incident of crying for a lover appears in other singing games, as, for instance, in “[Poor Mary].” Especially may this be considered the process which has been going on when it is seen that “choosing” is an actual incident of the game, even in those cases where “crying” has replaced the kneeling. The choosing incident also assumes two forms, namely, with respect to “east and west” in twenty-two versions, and “best and worst” in nine versions. Now, the expression, “for better for worse,” is an old marriage formula preserved in the vernacular portion of the ancient English marriage service (see Palgrave, English Commonwealth, ii., p. cxxxvi.); and I cannot but think that we have the same formula in this game, especially as the final admonition in nearly all the versions is to choose “the one loved best.” Following upon this comes the very general marriage formula noted so frequently in these games. It is slightly varied in some versions, and is replaced by a different formula, but one that also appears in other games, in two or three versions. One feature is very noticeable in the less common versions of this game, viz., the assumption of the marriage being connected with the birth of children, and the indulgences of the lovers, as in the Tong and Scottish versions [xxxii.], [xxxiii.], and [xxxiv.]

(e) In considering the probable origin of the game, the first thing will be to ascertain as far as possible what ideas the words are intended to convey. Taking note of the results of the analysis, so far as they show the corruptions which have taken place in the words, it seems clear that though it is not possible to restore the original words, their original meaning is still preserved. This is, that they accompanied the performance of a marriage ceremony, and that a chief feature of this ceremony was connected with some form of water-worship, or some rite in which water played a chief part. Now it has been noted before that the games of children have preserved, by adaptation, the marriage ceremony of ancient times (e.g., “[Merry ma Tansa],” “[Nuts in May],” “[Poor Mary],” “[Round and Round the Village]”); but this is the first instance where such an important particularisation as that implied by water-worship qualifies the marriage ceremony. It is therefore necessary to see what this exactly means. Mr. Hartland, in his Perseus (i. 167-9), draws attention to the general significance of the water ceremonial in marriage customs, and Mr. F. B. Jevons, in his introduction to Plutarch’s Romane Questions, and in the Transactions of the Folk-lore Congress, 1891, deals with the subject in reference to the origin of custom obtaining among both Aryan and non-Aryan speaking people. In this connection an important consideration arises. The Esthonian brides, on the morning after the wedding, are taken to make offerings to the water spirit, and they throw offerings into the spring (or a vessel of water), overturn a vessel of water in the house, and sprinkle their bridegrooms with water. The Hindoo offerings of the bride were cast into a water vessel, and the bride sprinkles the court of the new house with water by way of exorcism, and also sprinkles the bridegroom (Jevons, loc. cit., p. 345). Here the parallel between the non-Aryan Esthonian custom and the Aryan Hindoo custom is very close, and it is a part of Mr. Jevons’ argument that, among the Teutons, with whom alone of Aryan speaking peoples the Esthonians came into contact, the custom was limited to the bride simply stepping over a vessel of water. There is certainly something a great deal more than the parallel to the Teutonic custom in the game of “Sally, Sally Water,” and as it equates more nearly to Hindoo and Esthonian custom, the question is, Does it help Mr. Jevons in the important point he raises? I think it does. A custom is very low down among the strata of survivals when it is only to be recognised as part of a children’s singing game, and the proposition it suggests is that children have preserved more of the old custom than was preserved by the people who adopted a portion of it into their marriage ceremony. A custom so treated must be older than the marriage ceremony with which it thus came into contact, and if this is a true conclusion, we have in this children’s game a relic of the pre-Celtic peoples of these islands—a relic therefore going back many centuries for its origin, and which is of inestimable service in discussing some important problems of the ethnic significance of folk-lore. These conclusions are entirely derived from the significant position which this game occupies in relation to Esthonian (non-Aryan) and to Teutonic (Aryan) marriage customs respectively, and therefore it is of considerable importance to note that it entirely fits in with the conclusion which my husband has drawn as to the non-Aryan origin of water-worship (see Gomme’s Ethnology of Folk-lore, pp. 79-105).

There is, however, something further which seems to bring this game into line with non-Aryan marriage customs. The marriage signified by the game is acknowledged and sanctioned by the presence of witnesses; is made between two people who choose each other without any form of compulsion; is accompanied by blessings upon the young couple and prognostications of the birth of children. These points show that the marriage ceremony belongs to a time when the object of the union was to have children, and when its duration was not necessarily for life. It is curious to note that water worship is distinctly connected with the desire to have children (Proc. Roy. Irish Acad., 3rd ser., ii. 9); and that the idea of the temporary character of the marriage status of the lower classes of the people is still extant I have certain evidence of. Early in November of 1895, a man tried for bigamy gave as his defence that he thought his marriage was ended with his first wife, as he had been away seven years. It is a frequently told story. A year and a day and seven years are the two periods for which the popular mind regards marriage binding. “I was faithful to him for seven years, and had more than my two children,” a woman said to me once, as if two children were the required or expected number to be born in that period. If there is a popular belief of this kind, it is strangely borne out by this game-rhyme. “First a girl, and then a boy,” may also be shown to be a result to be desired and prayed for, in the popular belief that a man’s cycle of life is not complete until he is the father of a daughter, who, in her turn, shall have a son. Miss Hawkins Dempster obtained evidence of such a belief from the lips of a man who considered he was entitled to marry another woman, as his wife had only borne him sons, and therefore his life was not (like hers) complete.

The free choice of both woman and man is opposed to the theory of our present marriage ceremony, where permission or authority to marry is only necessary for the woman, the man being able to do as he pleases. This is now regarded as a sign of women’s early subjection to the authority of men and their subordinate place in the household. But it does not follow that this was the relative position of men and women when a ceremony was first found needful and instituted. I am inclined to think it must have been, rather, the importance attached to the woman’s act of ratification, in the presence of witnesses, of her formal promise to bear children to a particular man. Marriage would then consist of contracts between two parties for the purpose of, and which actually resulted in, the birth of children; of concubinage, or the wife consenting to children being born to her husband by another woman in her stead, if she herself failed in this respect (such children being hers and her husband’s jointly); of marriage without ceremony or set purpose, resulting from young people being thrown together at feast times, gathering in of harvests, &c., which might or might not result in the birth of children. These conditions of the marriage rite are at variance with what we know of the Aryan marriage generally and its results; and that they flow from the customs preserved in the game under consideration is further proof of the origin of the game from a marriage rite of the pre-Celtic people of these islands. The “kissing together” of the married couple is the token to the witnesses of their mutual consent to the contract.

Attention has already been directed to the fact that parts of the formula preserved in this game are also found in other games, and it may possibly be assumed therefrom that the same origin must be given to these games as to “Sally Water.” The objection to such a conclusion is mainly that it is impossible to decide to which game the popular marriage formula originally belonged, and from which it has been borrowed by the other games. Seeing how exactly it fits the circumstances of “Sally Water,” it might not be too much to suggest that it rightly belongs to this game. Another point to be noted is that the tune to which the words of the marriage formula are sung is always the same, irrespective of that to which the previous verses are sung, and this rule obtains in all those games in which this formula appears—a further proof of the antiquity of the formula as an outcome of the early marriage ceremony.[Addendum]