The text known in English as Winnie-the-Pooh occurs in dozens of different languages. Scholars have long debated the question of what was the original language of composition. One of the most popular hypotheses has been that the original text was written in English. The present paper will use textual evidence to demonstrate the impossibility of that hypothesis and to suggest a more likely candidate.
Consider the following lines from the beginning of Chapter I in the English-language version.
...here he is...ready to be introduced to you. Winnie-the-Pooh.
When I first heard his name, I said, just as you are going to say, "But I thought he was a boy."
"So did I," said Christopher Robin.
"Then you can't call him Winnie!"
"But you said---"
"He's Winnie-ther-Pooh. Don't you know what 'ther' means?"
"Ah, yes, now I do," I said quickly; and I hope you do too, because it is all the explanation you are going to get.
Readers unfamiliar with British English may need to be informed that ther is intended to be a representation of a stressed form of the English definite article. Orthographic word-final r is silent in British English but frequently indicates lengthening and stress on the preceding vowel.
Having explained that element of this rather confusing text, we are left with a less malleable problem. In what way is the word the meant to be understood as an indication of masculine gender? It simply makes no sense; therefore, the narrator's reaction, "Ah, yes, now I do", likewise is incomprehensible in the English version.
Consider, however, the same text in the Latin version:
...nunc ipse...vobis ostentari paratus. Winnie ille Pu.
Nomen audiens primum, sicut vos dicturi estis, atiam ego dixi, 'Credidi eum puerum esse.'
'Ego quoque,' dixit Christophorus Robinus.
'At non potes eum "Winnie" vocare!'
'Est Winnie ille Pu. Num nescis, quid "ille" significat?'
'Scilicet nunc scio,' cito subiunxi; et spero vos etiam scire, quia nullam nisi hanc explicationem accipietis.
To those who know Latin, and who therefore do know 'quid "ille" significat', the Latin text is entirely perspicuous. Ille is a 3rd sg. masc. distal demonstrative pronoun; it is the source of both the masc. sg. definite article and the 3rd sg. masc. personal pronoun in most Romance tongues; e.g. Spanish el and él, French le and il. In Latin, therefore, adding the word ille after Winnie is plainly a way of indicating that, contrary to the seeming feminineness of the name Winnie, the bear in question is in fact male.
One conclusion based on our textual comparison is inescapable. Winnie-the-Pooh was not originally written in English. Rather, it was rendered into English by a somewhat inept translator, who simply did a morpheme-by-morpheme rendering of the bear's name.
It is easier to state that English was not the original language of the text than it is to determine what was that original language. Latin is certainly a possibility. The text makes sense in Latin, and a text originally written in Classical Latin would be a natural candidate for eventual translation into European vernaculars. We might also point out that it would be strange for anyone in the past few hundred years to make a translation of such a text into Latin, since there is hardly anyone who could not read it more easily in a vernacular version.
However, a translation into Classical Latin might well have been made in medieval times, as a way of making more accessible a text which existed only in a single vernacular. Or, a text originally written in late Vulgar Latin or early common Romance could have been "spruced up", that is, edited to adhere to Classical diction, by some medieval redactor. Again, the text quoted above offers evidence relevant to the question. It would be quite odd, in Classical Latin, to call someone "Winnie ille Pu". One can hardly imagine Vergil referring to Octavian as "Gaius Julius ille Augustus" (note that the Pooh/Pu seems, like Augustus, to have originated as a title but quickly developed into a name). Ille was simply not a definite article in Classical Latin--but it developed into one by late Vulgar Latin, and is preserved as such in the various Romance languages. It therefore seems most probably that the Latin text is a translation from some Romance language, and that the Latin translator faced the same difficulty as the English one, namely, how to translate an original masculine singular definite article into a language which does not have masculine singular definite articles. The English translator chose to use a generic definite article, thus making the English translation grammatical, but semantically confusing. The Latin translator chose to use a masculine singular demonstrative pronoun, thus producing a grammatically non-standard text which, however, still makes semantic sense in context.
There is further evidence against Latin as the original textual language in the name "Winnie". "Winnie" does not look like a Classical Latin feminine noun. If the text had originally been in Latin, the bear's name would presumably have been "Vinnia". It is interesting that the Latin translator did not, in fact, choose to substitute "Vinnia" for "Winnie" in the course of Latinizing the text, since names frequently were Latinized in Latin translations. In any case, it is fortunate for us that the original form "Winnie" was preserved even in the Latin text, since it provides further evidence against the initially-attractive original Latin hypothesis.
Moreover, the name "Winnie" gives us a solid basis for determining what was the original compositional tongue, In most Romance languages, the final -a of first declension nouns has been preserved. In Gallo-Romance, however, it reduced to schwa and was spelled as -e. Moreover, while modern French only has orthographic w in borrowed words, particularly names, w was a regular element of the orthography of Gallo-Romance and Old French.
It therefore seems most likely that the original language of Winnie-the-Pooh was Gallo-Romance or Old French. Whether the former or the latter depends, in part, on whether one regards the adventures of Pooh and friends as more similar to a Gallo-Romance saint's life or to an Old French chanson de geste or roman d'aventure. To the present author's mind, the third genre is obviously the correct choice. The Pooh cycle is very similar in both form and subject to the Arthurian prose romances which began to be written in 13th-century France. Christopher Robin is a somewhat passive ruler figure, not unlike the Arthur of the romances. Pooh is his most beloved servitor and the active hero of most adventures, just like Lancelot, though of course Pooh never betrays Christopher Robin. Pooh has a variety of adventures, sometimes alone, sometimes in the company of one or more fellow members of the knightly circle. And we should also note that other French prose romances were very widely translated into other languages. The fact that the original Old French version of Winnie-the-Pooh has been lost should not lead us to doubt that it must have existed.
By way of an afterword, it should be noted that the theory advanced in this paper does not in any way imply that the Winnie-the-Pooh stories themselves, as opposed to the text known as Winnie-the-Pooh, are of French origin. Medieval French romancers wrote about Alexander the Great (a Greco-Macedonian), Aeneas (a Trojan/Roman), Arthur and Tristan (both British), but rarely about Frenchmen. Nor does Pooh seem to be particularly Gallic in personality. Most likely the Pooh stories originated as oral literature in some country other than France. A French writer became familiar with them and used them to create the roman d'aventure which is so popular today.