Broad and Narrow Transcription
Chapter Objectives
By the end of this chapter, you will be able to:
- Identify the purpose and assumptions behind a given transcription.
- Distinguish phonemic, broad phonetic, and narrow phonetic transcription.
- Select the consequential details a transcription task actually requires.
- Explain why narrow transcription is not automatically the better choice.
- Recognize alternative, equally defensible analyses of the same data.
- Create parallel representations of the same verified form at different levels of detail.
An Observable Problem
Chapter 23 compared a citation form written between slashes with a connected realization written between square brackets, and treated that difference mostly as a matter of context: careful, isolated speech versus fluent, connected speech. But even a single, carefully spoken word can be transcribed in more than one legitimate way. Consider the English word "little" in a rhotic-less General American-adjacent pronunciation. One transcriber might write /ˈlɪtəl/. Another, working from the same recording, might write [ˈlɪɾɫ̩], showing a tapped medial consonant, a syllabic lateral, and velarization. Neither transcriber has necessarily made an error. They have answered different questions.
This chapter asks three questions:
- What is actually being claimed by a transcription written between slashes, compared with one written between square brackets at increasing levels of detail?
- How does a transcriber decide, for a specific task, how much detail is worth including?
- Why is it a mistake to treat the narrowest possible transcription as the automatically correct or automatically best one?
Three Levels, One Utterance
Purpose: show the same verified form at three levels before naming any rule, so the difference is heard and seen before it is defined.
A single verified utterance can be represented at more than one level of abstraction. This chapter uses three working levels: phonemic transcription, which represents only the contrastive categories relevant to an analysis; broad phonetic transcription, which represents a defined, pedagogically useful degree of phonetic detail without attempting to capture every observable feature; and narrow phonetic transcription, which adds further observed or stipulated detail beyond the broad level.
Each level answers a legitimate, different question about the same recording. None of the three forms is a corrected version of another.
Purpose Drives the Level
Purpose: connect level selection to the purpose-dependent transcription principle established throughout the book, and to the phone-phoneme-allophone framework from Chapter 5.
A phonemic transcription represents the categories that contrast for a stated variety and analysis: swap one phoneme for another and the meaning of the word changes, or at least could change, for at least one pair of words in the language. A phonemic transcription deliberately omits predictable, allophonic detail, because that detail does not change which contrastive category a listener perceives, a distinction developed fully in Chapter 5.
A broad phonetic transcription is not simply "less careful" phonemic work. It is a phonetic transcription, written in square brackets, that a project or task has defined to include a specific, limited set of predictable detail, chosen because that detail is pedagogically or analytically useful at the stated level. A narrow phonetic transcription extends that same square-bracketed representation with additional detail, again chosen because a specific task requires it, not because more detail is inherently more accurate.
Before evaluating whether a transcription is "right," ask what question it was written to answer. A phonemic transcription answers "what contrastive categories does this speaker use here?" A broad phonetic transcription answers "what does this sound like at a level of detail useful for this task?" A narrow phonetic transcription answers "what additional, specific phonetic properties are observable or relevant here?" A transcription cannot be judged accurate or inaccurate without knowing which question it was answering.
Selecting Consequential Detail
Purpose: give learners a working procedure for deciding what belongs in a given transcription.
Not every phonetic detail available in a recording is worth transcribing. A useful working question is: does this detail matter for the stated purpose of the transcription? Detail is consequential when it is contrastive for the variety in question, when the task has explicitly requested it, when it is necessary to explain a specific pattern under discussion, or when omitting it would mislead a reader about what was actually said. Detail is not automatically consequential merely because it is audible, measurable, or interesting.
Error: Adding every diacritic a learner has ever seen to a transcription, regardless of whether the task requires it, in order to appear thorough.
Correction: Every diacritic in a transcription should be traceable to a stated purpose or observed evidence, consistent with the diacritic-selection principle introduced in Chapter 25. Unsupported or unrequested detail is overtranscription, not rigor.
Why Narrow Is Not Automatically Better
Purpose: directly counter the common beginner assumption that more detail always means a better transcription.
It is tempting to assume that a narrow phonetic transcription is simply a more accurate version of a broad or phonemic one, and that a careful transcriber should always prefer the most detailed representation available. This assumption is mistaken for several reasons.
First, a phonemic transcription is not an incomplete phonetic transcription; it is a different kind of claim, about contrastive categories rather than physical detail, and adding phonetic detail to it does not make the phonemic claim more correct. Second, a narrow transcription that is not supported by genuine evidence, such as an instrumental measurement or a careful, controlled listening judgment, can overstate the transcriber's actual certainty. Third, excessive narrow detail can obscure the pattern a reader actually needs to see, burying a consequential contrast under a large number of inconsequential ones. Fourth, narrow transcriptions are harder to produce reliably and harder to compare across transcribers, so a task that does not require that level of detail incurs unnecessary cost and risk of error for no analytical benefit.
A narrow phonetic transcription is only as good as the evidence behind each added symbol. Acoustic evidence, discussed further in Chapter 28, can support some narrow claims more reliably than unaided listening alone. A transcriber who adds detail without a clear evidentiary basis is not being more careful; they are making an unsupported claim that happens to look precise.
Recognizing Defensible Alternatives
Purpose: prepare learners to accept more than one correct answer when the underlying analysis genuinely allows it.
Two trained transcribers working from the same recording can produce different, equally defensible transcriptions at the same stated level, because phonetic and phonological analysis sometimes leaves more than one reasonable option. This can happen when a boundary case falls between two categories, when a language's own descriptive tradition offers more than one accepted convention, or when a predictable process could reasonably be represented either by showing its output or by leaving it to a general rule stated once in a transcription key.
Recognizing a defensible alternative is not the same as accepting an arbitrary or unsupported choice. A defensible alternative must still be internally consistent, stated explicitly in a transcription key, and applied the same way throughout a single analysis.
Levels Across Languages
The phonemic-versus-phonetic distinction applies to every language with a describable sound system, not only to English. A language with a small phoneme inventory can still show extensive allophonic variation that only a broad or narrow phonetic transcription would capture, and a language with a large phoneme inventory does not automatically require narrower transcription than one with a small inventory. The amount of allophonic detail worth representing depends on the specific sound system and the specific task, not on the size of the phoneme inventory alone.
Analysis Exercises
These exercises apply the three-level distinction and the purpose-driven selection procedure developed in this chapter. State the reference variety and requested level for each answer.
Exercise A — Matching Level to Purpose
For each stated purpose, decide whether a phonemic, broad phonetic, or narrow phonetic transcription is most appropriate, and briefly justify your choice.
- A dictionary pronunciation key intended to show contrastive categories only.
- A beginner's guide showing learners the flapped consonant they are likely to hear in casual American speech.
- A laboratory report documenting precise voice onset time and secondary articulation for a research study.
Exercise B — Removing Unjustified Detail
The following narrow transcription was produced for a task that only required a phonemic representation. Rewrite it at the phonemic level and explain which details were removed and why.
[ˈkʰæmʰpʰɪŋ]
Exercise C — Two Defensible Analyses
Explain why two trained transcribers might reasonably disagree about whether to write a word-final syllabic consonant as a syllabic segment or as a vowel-plus-consonant sequence, and state what a transcription key should do to resolve the ambiguity for a reader.
Common Mistakes
Error: Treating phonemic transcription as an inferior, less accurate version of phonetic transcription.
Correction: Phonemic transcription makes a different kind of claim, about contrastive categories, and is not made more correct by adding phonetic detail it was never designed to represent.
Error: Assuming a single, universal broad transcription exists for a word, independent of the analysis or variety chosen.
Correction: A broad transcription reflects a defined level and a stated variety. Different projects can define broad transcription differently, provided each defines and applies its convention consistently.
Error: Silently mixing phonemic and phonetic detail within the same transcription, or mixing conventions from two different varieties.
Correction: Choose one level and one variety for a given transcription, mark the level with the correct delimiters, and state the convention explicitly if it could otherwise be ambiguous.
Chapter Review
Part A — The Three Levels
1. Define phonemic, broad phonetic, and narrow phonetic transcription in your own words, and state which delimiters mark each.
2. Explain what question each level of transcription is answering.
Part B — Selecting Detail
1. Explain the working procedure this chapter recommends for deciding whether a phonetic detail belongs in a given transcription.
2. Explain why narrow transcription is not automatically the best choice for every task.
Part C — Alternatives and Evidence
1. Explain what makes an alternative analysis defensible rather than merely arbitrary.
2. Explain what kind of evidence should support an added narrow-transcription detail.
Cumulative Review
1. From Chapter 5: Explain how this chapter's three levels build on the phone-phoneme-allophone distinction introduced there.
2. From Chapter 23: Explain how the citation-versus-connected comparison in Chapter 23 previewed this chapter's broad-versus-narrow distinction.
3. Preview for Chapter 25: This chapter selected consequential detail in general terms. Predict how Chapter 25 will make individual diacritic choices more precise.
Self-Check
Before moving to Chapter 25, confirm that you can do the following without looking at your notes:
- State the purpose and assumptions behind a given transcription.
- Distinguish phonemic, broad phonetic, and narrow phonetic transcription, and mark each with the correct delimiters.
- Decide whether a specific phonetic detail is consequential for a stated task.
- Explain why narrow transcription is not automatically superior to broad or phonemic transcription.
- Recognize when two transcriptions can both be defensible.
- Produce parallel phonemic, broad, and narrow representations of the same verified form.
If any of these tasks are difficult, review the relevant section before reading Chapter 25. Chapter 25 consolidates the principled use of individual diacritics and suprasegmental symbols, building directly on this chapter's detail-selection procedure.
Answer Key (click to expand)
Analysis Exercises — Exercise A
1. Phonemic; a dictionary pronunciation key intended to show contrastive categories only does not require predictable phonetic detail. 2. Broad phonetic; showing a beginner what they are likely to actually hear requires the predictable, pedagogically useful detail that a broad level is designed to represent. 3. Narrow phonetic; a laboratory report documenting precise voice onset time and secondary articulation requires the additional, evidence-supported detail that only a narrow level provides.
Analysis Exercises — Exercise B
/ˈkæmpɪŋ/. The aspiration diacritics on the initial and medial stops were removed because aspiration is predictable in this position for this variety and is not part of a phonemic-level contrast; the phonemic level represents only the contrastive segments required to distinguish this word from others in the language.
Analysis Exercises — Exercise C
Two transcribers might disagree because syllabicity at a word boundary can be a genuine borderline case, with some tokens showing a clear vocalic release and others showing a fully consonantal, syllable-bearing segment, and because different established conventions treat this pattern differently. A transcription key should state explicitly which convention is being used and apply it consistently throughout the analysis, so a reader can interpret every instance correctly even if they would have made a different choice themselves.
Chapter Review — Part A
1. Phonemic transcription, marked with slashes, represents only contrastive categories; broad phonetic transcription, marked with square brackets, represents a defined, limited degree of phonetic detail; narrow phonetic transcription, also marked with square brackets, adds further observed or stipulated detail beyond the broad level. 2. Phonemic transcription answers what contrastive categories a speaker uses; broad phonetic transcription answers what a form sounds like at a level of detail useful for a stated task; narrow phonetic transcription answers what additional specific phonetic properties are observable or relevant.
Chapter Review — Part B
1. Ask whether a detail is contrastive for the variety in question, explicitly requested by the task, necessary to explain a pattern under discussion, or necessary to avoid misleading a reader; detail that meets none of these conditions should normally be omitted. 2. Narrow transcription is not automatically best because a phonemic transcription answers a different kind of question rather than a less complete one, because unsupported detail overstates certainty, because excessive detail can obscure the pattern a reader needs to see, and because narrow transcription is harder to produce and compare reliably.
Chapter Review — Part C
1. An alternative is defensible when it remains internally consistent, is stated explicitly in a transcription key, and is applied the same way throughout a single analysis, rather than being an arbitrary, inconsistent, or unstated choice. 2. Added narrow-transcription detail should be supported by acoustic evidence or a careful, controlled listening judgment, not added merely because it seems plausible or thorough.
Cumulative Review
1. Chapter 5 distinguished phones, the physical realizations of speech, from phonemes, the contrastive categories a language treats as functionally distinct, and allophones, the predictable variants of a phoneme; this chapter's phonemic level corresponds to the phoneme level of that framework, while broad and narrow phonetic transcription represent increasing amounts of allophonic and phonetic detail. 2. Chapter 23 already used slashes for a citation-form phonemic representation and square brackets for a connected phonetic realization, which is a specific instance of the general phonemic-versus-phonetic distinction that this chapter now names and extends into a full three-level system. 3. Chapter 25 is likely to take the general detail-selection procedure developed here and apply it symbol by symbol, teaching learners exactly which diacritics are available, what each one means, and when a specific diacritic is the consequential detail a task requires.
Looking Ahead
What You Can Now Do
You can now identify the purpose and assumptions behind a transcription, distinguish phonemic, broad phonetic, and narrow phonetic representation, select the consequential detail a task actually requires, explain why narrow transcription is not automatically superior, recognize defensible alternative analyses, and produce parallel representations of the same verified form at different levels of detail. Together with Chapter 23, you now understand both how context changes a realization and how purpose changes a transcription, the two central ideas that Part V builds on.