(I've taken a look into this matter - and the notes that I've been making during the process have piled up considerably. So let me share them here for the potential future reference; maybe someone will find them interesting.)
The Entu, Ensi, Enta Declensions (EEED) is a chart filling one side of a sheet of the so called Oxford paper, a medium that Tolkien started being supplied with in 1924.
The other side contains earlier notes on G. Chaucer and his use of Middle English (cf. VT 39/7). Interestingly, in 1922 Tolkien indeed started working on an extensive edition of Chaucer's texts - and kept being busy with it until mid 1925 and then again in 1930-1932 (The J. R. R. Tolkien Companion and Guide - Chronology, p. 118-130,155-158).
So it might be concluded that this particular sheet of paper became available for its secondary use not earlier than 1924 and possibly not later than 1932.
Thanks to a very contributive observation by Paul Strack, we have recently learned that the case endings given in EEED are closely resemblant of those in Quenya Declensions Version 3 (PE 16/127-128) - even to the point of being practically identical.
Quenya Declensions Version 1-3 (PE 16/124-128) were, together with the Qenya Word-lists (PE 16/132-148), created in preparation for a set of Quenya poems subsequently included in the lecture Secret Vice (cf. PE 16/105,116,129-131). The lecture was given in 1931 and most of the poems may not be much older (cf. PE 16/53). So V1-3 and QWL - and consequently also EEED - are likely to date either to 1931, or to a year or a few years earlier.
E. Kloczko claims (Elfling messages #02673, #02674, #02742 from May-June 2000) that EEED is a fragment of a larger complex, of which another piece - a Qenya conjugations chart (QCC) - is also found in Marquette. (The so-far blank versos of both of them have reportedly been later, in the 1940s, used for LotR drafts.) Kloczko isn't renowned for his reliability, nevertheless he has seen the material and I haven't, so I'm taking his word on this and accepting that both documents probably do resemble each other in some significant manner.
Now, it has been proposed in 1994 by the editors of EEED (VT 39/9-18) that entu, ensi and enta are demonstrative pronouns - either varying in their spatial reference (like "this here" vs. "that over there"), or inflected for gender.
Some relevant materials have been published since then:
On one hand, QCC (in this case Elfling #02742) gives a trio of pronouns inqi "this (by me)", este "this/that (by you)" and enta "that (neither by me nor by you)", the third one of them identical with one of those in EEED. However, the other ones are significantly different.
On the other hand, personal endings shown in early Qenya Conjugations (in this case chiefly Version 1, PE 16/124) appear to be basically m. to or do, f. se or te and n. ta or da in singular and tur/sur, tir/sir and tar in dual - closely resemblant of the stem forms ento, ense, enta (singular) & assut, assit, assat (dual) in EEED.
Furthermore, QCC reportedly gives personal endings almost identical to those in Qenya Conjugations (Version 2-3, PE 16/125-128).
And finally, Qenya Conjugations was written on a Leeds examination paper, which Tolkien couldn't have obtained later than 1925 (VT 16/116) - and although he soon started keeping this material together with his later Quenya Declensions Version 5-6, created in the mid 1930s, years after the Secret Vice (ibidem), this might have been simply because there was no newer material on conjugations (while declensions had undergone significant changes and therefore have been written about over and over again). The Quenya Conjugations are at any rate in accordance with the language used for the Secret Vice poems (ibidem).
All in all, whatever do the words in EEED mean, the evidence points towards this document being a close contemporary if not a complement to the Qenya Conjugations Ver. 1-3 and to the Quenya Declensions Ver. 1-3, while at the same time indeed being in a close relation to QCC. Consequently, all these materials form an intertwined set created between 1925 and 1931, possibly rather later than earlier within this time span.
After he had finished the Secret Vice (and also abandoned other works, the Chaucer edition included), Tolkien is known to have started two successive versions of a new complete Qenya grammar - one in early 1930s (PE 21/vi; let's call it Grammar 1) and the other one, a replacement for the former, some years later, chiefly developed during 1936 (PE 21/ix-x; Grammar 2). Both were abandoned (in favour of the newly started Tengwesta Qenderinwa and the Etymologies) by 1937 - and in both cases, the first sections of the texts, treatises of the (historical) phonology, have now been lost (apparently due to their replacement by the Tengwesta). They only survive in the form of a few scattered fragments on versos of various later works (cf. ibidem).
Apart from this historical phonology, Tolkien doesn't seem to have gotten further then the declensions of nouns - namely the practically finished text of the Declension of Nouns (PE 21/1-41) for Grammar 1 and only the charts known as Version 5-6 (plus possibly the Bodleian Declensions and the Note on Final Consonants, altogether PE 21/42-61 and VT 28/8) for Grammar 2. As mentioned above, no section on verbs (or other parts of speech) is known to have existed and Tolkien evidently kept using his conjugations charts from before 1931.
However, this also means that he apparently considered the pre-1931 verb forms still more or less valid during the mid 1930s. And that opens the possibility that QCC, rather than being a pendant to the older Quenya Declensions, belongs with the scattered fragments of unfinished Gramamar 1 or Grammar 2. This would invalidate E. Kloczko's statement about the relation to EEED, but on the other hand, it would give the abovementioned pronominal forms inqi and este a later datation, leaving entu or ensi free to be a conceptual precursor to some of them.
Naturally, a visit to the Marquette University library would probably provide an answer to this.