How an Index Preserves Scholarly Uncertainty

An index heading can change an argument without adding a single word. Remove “possible” from “possible influence of metre,” and a tentative explanation begins to look like a finding. Remove “proposed” from a historical account, and a view the author discusses may appear to be a view the author accepts.

For authors, indexers and editors, this is a compression problem: the index must shorten the discussion while preserving what is claimed, by whom, and with what confidence. A correct page number cannot repair a heading that promises the wrong argument.

Preserve the claim, not every hedge

Suppose a chapter says: “Metre may account for the longer form in these poems, although the surviving examples do not establish a general rule.” A heading such as “Verb forms—metrical conditioning” loses both the uncertainty and the limit to poetry. “Verb forms—possible metrical influence in poetry” preserves the useful distinction without reproducing the whole sentence.

Not every cautious phrase needs an index equivalent. “It seems useful to begin with the inscriptions” says how the author will proceed. “The inscription may preserve a borrowed form” qualifies the interpretation itself. The second qualification belongs in the heading when dropping it would change the reader’s expectation.

Uncertainty has more than one meaning

These five sample passages call for different wording. Treating all of them as “debate” or “uncertainty” conceals the result the reader is looking for.

• Proposed: “An earlier account explains the long form as a politeness marker.” Use “Long forms—politeness explanation, proposed.” Reporting a proposal does not show that the chapter adopts it.
• Possible: “Rhythm may influence the choice in verse.” Use “Long forms—possible rhythmic influence in verse.” Keep the limited setting as well as the tentative status.
• Questioned: “Written spellings have been treated as records of speech, but literary imitation complicates that inference.” Use “Written forms—as evidence of speech, questioned.” This identifies the inference under scrutiny.
• Inconclusive: “Too few examples survive to decide whether neighbouring words affect the choice.” Use “Form choice—collocational evidence, inconclusive.” Insufficient evidence does not establish that the proposed influence is absent.
• Rejected: “The chapter’s comparison rules out politeness as an explanation of this alternation.” Use “Form choice—politeness explanation, rejected.” Keep the rejection within the specific alternation; it says nothing about politeness elsewhere.

Choose the qualification from the passage, not from a fixed vocabulary. If the author only reports counterexamples, “counterevidence” may be more accurate than “rejected.” If a proposal is later accepted, the index should also give access to the supporting discussion. A qualifier is a compact statement about the argument, not a confidence rating supplied by the indexer.

Separate claims that answer different questions

Steven E. Fassberg’s “III-y Imperatives in Ancient Hebrew” illustrates a useful distinction. Its history of scholarship reports an earlier association between a vowel pattern and Aramaic; its corpus discussion considers an Aramaic analysis of an imperative in the Bar Kosiba letters. One concerns an explanation of a pattern, the other the linguistic classification of particular evidence. This is my indexing interpretation of those discussions, not a reproduction of a published index.

For an invented parallel, imagine that page 42 proposes influence from Language B on a vowel pattern, while page 87 considers whether one letter’s greeting belongs to Language B at all. “Language B—influence and identification, 42, 87” combines two questions. More useful subentries would be:

• Language B—proposed influence on vowel pattern, 42
• Language B—possible language of letter greeting, 87

Assign each locator to the claim it supports. Copying both pages into both subentries preserves the appearance of coverage but loses the distinction the split was meant to recover. The letter collection could also offer access to the greeting’s disputed classification if readers are likely to seek it there.

A split is not required merely because a heading names two things. “Short and long forms—comparison” can describe one sustained comparison. Separate subentries earn their place when they serve different reader questions, distinguish competing accounts, or preserve different evidentiary outcomes.

Read the conclusion before settling the heading

An early passage may introduce an attractive hypothesis; a later passage may narrow or reject it. Indexing only the introduction risks freezing the argument before its conclusion. Check the later discussion and synthesis before deciding whether the heading needs “proposed,” “supported,” “questioned,” or a more specific description.

Also preserve the object of the conclusion. “The surviving written forms do not establish spoken usage” cannot become “Spoken usage—absence.” The first limits an inference from evidence; the second asserts a historical fact. An index should help the reader find that limit, not silently cross it.

A short review before approval

• ✓ Whose claim is this: the author’s, a cited scholar’s, or an alternative being tested?
• ✓ What does the source conclude, and does a later passage change the earlier proposal?
• ✓ Does the heading retain any necessary limit of corpus, period, form or context?
• ✓ Does each page support the complete heading and subentry, including its qualification?
• ✓ Would merging two entries conceal different claims or different outcomes?

Read the finished heading as a promise to the reader. Keep the shortest wording that preserves the source’s claim and its limits.