(from the OED)
“Of difficult etymology.
“After the solitary (but app. unquestionable) occurrence in Shoreham's Poems c1315 the word first appears in 1531, and soon afterwards is very common; prob. it had survived only as a technical term of some rustic sport, and so failed to be recorded in literature.
“Shoreham's form gol, riming with y-hol, suggests (though it does not necessarily imply) descent from an OE. *gál. (This would be quite certain if the word could be positively identified with the gale found in Cursor M. 8710 (Cott.): ‘O þis quick þai bath wald be Moder‥And aither wald þai haf it hale, Bot þai mai neuer com to þat gale’. But it is not unlikely that in this passage gale, obviously chosen for the sake of rime, is merely a forced use of GALE n.2 in the sense of ‘joy’. Of the existence of an OE. n. *gál, with the sense ‘obstacle, barrier’, some indirect evidence is afforded by the apparent derivatives gælan, ágælan, to hinder, delay. The transition from the sense of ‘barrier’ to that of ‘boundary’ (sense 1) is easy, and the further sense-development is parallel to that of L. meta, and of DOOL n.2; in view of the history of the latter word, Halliwell's alleged ‘Goale, a barrow or tumulus’, might be compared, if there were any ground for believing it to be genuine. But the absence of any record of OE. *gál or of its equivalent in any Teut. lang. (ON. geil, narrow passage, being too remote in sense) renders this etymology very insecure.
“The suggestion of Henshaw (in Skinner's Etymologicon 1671), accepted by all subsequent etymologists, that the word is an adoption of F. gaule, pole, stick, switch, has nothing to recommend it. There is no evidence that F. gaule ever meant ‘goal’, or that Eng. goal ever meant ‘pole’ or ‘switch’. Besides, for the form in Shoreham, the Fr. derivation appears to be phonologically inadmissible.
“A Welsh gâl is given by Davies Antiq. Ling. Brit. Dict. I. (1632) with Latin renderings (stadium, meta, statio) which would make it equivalent to Eng. goal. If this word be genuine (which seems to be doubtful) it must apparently be an early adoption from English; the suggestion in Fick Idg. Wb.2 II., that it represents an OCeltic *gaslâ stone, being phonologically untenable (Prof. Rhys). The current word in most parts of Wales for ‘goal’ with reference to games is col, mutated gol (gol), which prob. has obtained this meaning through its similarity of sound to the Eng. word. The Windhill dialect has a word pronounced (gol), explained as meaning ‘goal’, used in a certain game played with brass buttons (cf. GOG4); but its identity with this word is doubtful.”