The verb to go is irregular, and apart from be is the only suppletive verb in the English language.

Principal parts

The principal parts of the word are go, went, gone. Otherwise the modern English verb conjugates regularly. The irregularity of the principal parts results from the fact that they derive from two or possibly three different Indo-European roots.

The preterite (or 'simple past tense') is in no way etymologically related to go, for went comes from wendan in Old English, which is also the source of wend. Old English wendan and gān (the latter of which means go) did share semantic similarities, and their similar meanings can be seen in the fact that the sentence "I'm wending my way home", means "I'm going home."

Theories concerning the origin of gone are discussed below.

Origin of ēode

Old English didn't have the preterite went in any form, instead using the word ēode, a word which has not left any trace in modern English in any form. When one looks at ēode, in all its conjugated forms, it is not surprising to see all the –d's, for these are the familiar Germanic dental suffixes, establishing ēode as a preterite. The root itself, ēo, came from the unattested Proto-Germanic *ijjôm. The Gothic form of this root is iddja, but this form hasn't produced any other attested root words in the other Germanic languages. *Ijjôm was itself a past tense form of the Proto-Indo-European (PIE) root *yâ (go). Specifically, this root was either imperfect or aorist. (The aorist tense expressed momentary action in the past, while the imperfect, continual action in the past). *Yâ itself seems to have come from a PIE form *ei, î, and if this is correct, it would establish a link between the Old English Preterite for go and the Latin īre (go, pres inf.) (which is simply the î from *ei, î followed by a standard Latin infinitive ending, –re). The OED does not discuss this, but the 4th Edition of the American Heritage Dictionary does in its appendix of PIE stems, drawing heavily on Julius Pokorny's Indogermanisches etymologisches Wörterbuch (page 293). (īre is the source of many English words, words as disparate as introit, preterite, and ambition).

Development of a new preterite

Returning to the etymology of go, our now-familiar ēode became, in ME, variously ȝede, yede, and yode. By the 15th century in southern England, wende (wend) had become synonymous with go, but its infinitive and present tense forms had ceased to be in frequent use. With a waning, morphing preterite tense (yode), go was ripe to receive a new preterite—the preterite of wende, the familiar went. In Scotland and in the dialects of northern England, yede was also replaced, but by gaed, which was produced by adding a regular dental suffix to the regional variant of go. Went made it into standard English because southern England was to become the politically, culturally, and economically central region of England in modern British history. However, a writer of no less importance than Spencer used yede to mean go in some instances, with its preterite form of yode, but this was dialectical.

Etymology of wend

Wend (the source of go's current preterite) came from wendan. Wendan is thought, on the basis of numerous Germanic cognates, (particularly Gothic wandjan), to have come from the PIE root *wand. This root would be the preterite stem of windan. The relationship between windan and wendan needs to be briefly addressed.

Relationship between windan and wendan

The original form from which we get went is windan, which had wendan as a preterite stem, which in turn gave us went. Windan is not surprisingly the source of the modern verb wind (whose preterite and past participle is wound). The original preterite of windan was *wand-, and windan had a causative form, wendan (meaning "to cause to wind", or "to cause to become wound"). So, went is derived from wendan, which is itself derived from windan. Let us now investigate the etymology of windan.

Origins of windan

The Oxford English Dictionary's entry for wand simply states that words like wend, wind, wand, and wander all have a common PIE root, and that this root is related to the idea of turning. (Note that wand originally meant a supple switch, not a stiff rod, and is related to the word from which whip is derived.) The most important IE root (found in Pokorny 3. *er- 1152.) is treated in one of the American Heritage Dictionary's etymological indices under *wer-2. Though this root also carries with it the idea of turning, none of its English descendants are the words for which we are looking. Many turning-related words do come from *wer-2 (which Pokorny calls er-). For instance, we have wrist, wreath, writhe, (all of which involve turning), wring, wrench, and worm are only the most obvious descendants of this root. So, all we can say is that wind is derived from a similar PIE root to *wer-2.

The root *w- presupposed turning or motion, and was probably used both transitively and intransitively. Though originally wend meant to cause to wind (and the winding often being done in an intransitive sense), due to the similarity of these two words, they have been confused for at least a thousand years, and have thus influenced each other's developments. For much of their histories, wend and wind have had the sense of going, and thus it is not surprising that wend eventually came to have the sense of go. Winds past tense verb is winded.

Origins of the infinitive

*Ghê- is the PIE root from which go comes. It had the sense of "To release, let go; to be released; to go" (but in the middle voice). From *ghê, comes Old English gân (to go) and German gehen (which is relatively regular, compared to English go). Though the 1st person present indicative for go in Old English was gá, aside from an unsurprising shift from an a to an o, there has been little change in the infinitive form of this word for its entire history. It is rare for such a common word to undergo so few changes over such a long history.

Origins of the past participle

Gone is closely related to the now-obsolete verb, gang. Gang means "to walk" or "to go", (Scots: I'll gang nae mair tae yon hoose!) and is possibly the source of the past participles gone and German gegangen (which also means gone). According to this theory, the preterit of a form of gang eventually became past participles in English, German, and other related languages. The question arises of the relationship between gai-, the form responsible for present forms, and gang-. The OED describes three main theories:

• The two have no etymological connection, but have become similar in form because of their similar meanings.
• Gang- is a nasalized reduplication of gai-.
• The shorter gai- was created from gang- by analogy of stai- from stand- (the latter two relate to the verb "to stand").

Others have proposed a link between the Germanic forms and similar words in other Indo-European languages, but such theories have not attained general acceptance by the linguistic community.

Summary of the main Proto-Indo-European roots

Thus, we see that go (historically, anyway) is derived from at least 3 Proto-Indo-European roots: *ghê (from which we get go, and possibly gone), *ei, î, the source of ēode, and a root beginning in *w- from which we eventually get went, through windan and then wendan. We use three of the derivatives of these roots today, go, gone and went.

References

• The Online Etymology Dictionary
• The Oxford English Dictionary
• The Indogermanisches etymologisches Wörterbuch
• The American Heritage Dictionary's PIE roots index.