That’s the spirit: revising ‘spirit’, n. for OED3

The OED entry for spirit, n. consists of 163 senses, phrases, compounds, and so on, making it one of the larger entries in the dictionary, but by no means the largest. Also, despite its size, there is surprisingly little in spirit that is completely new – no new senses have been added to the entry, and only a small number of phrases and compounds (about which more will be said below). One might therefore suppose that the revision of this entry would be fairly uncomplicated: a lengthy (but not too lengthy), leisurely process consisting of little more than the addition of illustrative quotations from the twentieth and twenty-first centuries, some changes to definition text, and the occasional expansion of a sense or phrase that the first edition of OED mentioned in passing into a fully-fledged lexical item. In other words, a fairly cushy job.

How wrong could I be?

Pound for pound, spirit is easily the most awkward, frustrating, and downright difficult entry I’ve ever worked on. One reason for this is the subject matter: the supernatural is, unsurprisingly, not very easy to pin down in words, complicated by the fact that some senses are quite close to each other in meaning. The structuring of the entry (see below) presented another problem. Furthermore, with the exception of the more straightforward compounds, nothing in the entry was easy. Definitions were sweated over, talked through with colleagues, written and rewritten and re-re-written like so many miniature Hollywood scripts – always with the nagging suspicion that I or someone else might suddenly notice some incongruity or unaccounted-for shade of meaning that would upend everything. This article explores some of the difficulties that we encountered, and some other noteworthy features of the revised entry, in the hope that this will offer an interesting glimpse of the OED revision process.

Branching out

As mentioned above, the entry for spirit contains a large number of senses, phrases, and compounds (and a single derivative, spiritwise). The first edition of OED organized these into five top-level groupings, or ‘branches’, of semantically related senses (plus a sixth for compounds, which OED3 treats instead in a separate section). This sort of structure is typical of entries of this size; in fact, it would be very odd for a large entry not to be divided into branches, given that entries are normally large precisely because the word in question has developed in several different directions. However, in the case of spirit, much of this flourishing predates the first appearance of the word in English: Old French and Anglo-Norman esperit, espirit, etc. (the earlier forms of modern French ésprit) and its Latin equivalent spiritus had a wide range of meanings, which were imported wholesale into English,  especially from the Vulgate Bible. If you look at the entry you will notice, again and again, that the earliest evidence for a particular sense comes from the translations of the Vulgate into English made under the direction of John Wyclif in the late fourteenth century. Since the OED is a historical dictionary, the primary organizational principle within each entry is chronological, but when so many senses are dated to the same year, and when there are several significant strands of meaning that might be obscured in a strict chronological arrangement, the issue of structure becomes paramount.

OED1’s five branches can be summarized as follows:

• 1. An animating or vital principle; the soul; an incorporeal, supernatural, rational being, and extended uses.
• 2. The divine nature or essential power of God; a guiding or governing principle or power, and related senses.
• 3. The animating intelligence or sentient element of a person, and related senses.
• 4. Breath, and related senses; vital power; an intangible element in material things.
• 5. A volatile substance, especially one extracted by distillation.

It became clear that there would have to be some changes to this arrangement when I saw that OED1 had not followed a strict chronological sequence within these branches. There were good reasons for this: when the first edition of the OED was being worked on, there were far fewer resources at their disposal, particularly for the medieval and early modern period, than we have access to today, so they would occasionally organize their senses according to what they believed was the logical development of the word. OED3’s policy, however, is to organize entries chronologically – and, as we have seen, a great deal of the development seen in spirit n. took place long before it was first used in English, so we should not expect the senses to appear in ‘logical’ order.

The main chronological problems were caused by the small amount of evidence which pre-dated the Wycliffite Bible: in particular, the earliest evidence for the sense ‘a supernatural being’ was older than that for the sense ‘the soul’ – yet ‘an animating or vital principle’ and ‘the soul’ were plainly closer in meaning to each other than either was to ‘a supernatural being’, so I wanted to keep those senses close to each other. I was also not very happy that a branch beginning with ‘an animating or vital principle’ ended with a species of duck blessed with a spirit-like knack for disappearing and reappearing! So I experimented with moving all the senses relating to ‘a supernatural being’ (including the duck) into their own branch, which seemed to work well.  Further ruminating, and discussion with very patient and helpful colleagues, led to the realization that the ‘animating intelligence’ senses actually had quite a lot in common with the ‘animating principle’ ones, which led to these being merged into a single branch, although the datings of these senses meant that the two groups needed to be organized independently.

Conversely, I felt that the senses relating to the divine nature or essential power of God were sufficiently distinct from the ‘guiding or governing principle’ ones that these should be separated. OED1’s branch five, which was a bit of a grab-bag, was stripped back to consist solely of senses relating to breath, with the senses relating to a person’s faculties or qualities being added to the ‘animating intelligence’ group, while the ‘volatile substance’ branch became the repository of all the senses which related to actual material substances.

All of which led to the branch structure which has been adopted for OED3:

• 1. An animating or vital principle; the immaterial or sentient element of a person
  • An animating or vital principle; the soul; incorporeal or immaterial being.
  • The animating intelligence or sentient element of a person, and related senses.
• 2. An incorporeal, supernatural, rational being, and extended uses.
• 3. Breath, and related senses.
• 4. The divine nature or essential power of God.
• 5. A guiding or governing principle or power, and related senses.
• 6. A substance, essence, etc., formerly believed to animate or provide life; a volatile substance, esp. one extracted by distillation.

This structure isn’t perfect, but it constitutes a rather nice thumbnail sketch of the main strands of meaning, all of which were borrowed into English in the earliest period of the word’s history, while also conforming to OED3’s policies on chronological organization.

Moving material

As discussed above, there has been a lot of movement of material from one part of spirit, n. to another. Some material has also been moved into a new phrases section, such as the spirit is willing but the flesh is weak (the earliest example of which is found, yet again, in the Wycliffite Bible), while two compounds have been upgraded to become separate headwords in OED3: Holy Spirit, n. and spirit house, n. In the case of Holy Spirit, the compound’s cultural and religious significance were strong arguments in favour of giving it separate treatment, as was the fact that this would enable a fuller discussion of its etymology. In the case of spirit house, it was the discovery of new material which led to its being upgraded. This compound was not covered by OED1, but was included in the supplement published in 1986, and subsequently in OED2 in 1989, among a selection of undefined compounds. However, the quotation evidence included in OED2 clearly warranted a definition, since most of it related to a particular type of shrine in south-east Asia. Furthermore, our research uncovered two earlier senses of spirit house, both now rare but still in use, which related instead to alcoholic spirits: ‘a warehouse trading in alcoholic spirits, or in which alcoholic spirits are stored’ and ‘a tavern or public house licensed to sell alcoholic spirits’. The existence of three distinct and significant meanings for spirit house made it clear that this compound required an entry of its own.

What’s new

As mentioned earlier in this article, there are relatively few lexical items in spirit, n. which are new to this edition of the OED, and even fewer of these can be considered in any way new to the English language. A few phrases have been added, probably the most well-known of which is to enter (also get) into the spirit of, first attested in the late eighteenth century. Of the newly-added compounds, only spirit screen (1909) and spirit week (1923) date from the twentieth century, and the remainder are all first attested in the nineteenth. Most of these relate in some way to the supernatural, but an exception is the prosaic spirit vinegar (1865), which sticks in my mind because it wasn’t until very late in the revision process that I noticed it was missing! Fortunately by that point all the hard work had been done, and it presented no real problems in terms of research or definition, but the thought ‘will this ever end?’ definitely crossed my mind.

And finally

As you can imagine, when I finally finished work on spirit, n., I felt a great sense of relief. Now that it’s published, my main feeling is one of satisfaction: the entry is updated for the twenty-first century, as well as being, hopefully, more useful to the reader. But I daren’t look at it too closely, for fear that something will emerge to…

… wait for it…

… come back to haunt me.

Image by Tandem X Visuals on Unsplash

The opinions and other information contained in the OED blog posts and comments do not necessarily reflect the opinions or positions of Oxford University Press.