Transcription quality is directly limited by recording quality. Ensure your recordings are clear, with minimal background noise. Use a professional external microphone when possible, sample at 44.1 kHz / 16-bit minimum, and store files in a lossless or uncompressed format (WAV, FLAC) before beginning any transcription work.

The choices you make throughout this workflow — transcription system, level of detail, and prosodic notation — depend fundamentally on your research objectives and your intended audience. The level of phonetic detail, the choice of transcription system, and the extent of prosodic notation all vary according to who your transcription is for and what you intend to use it for. A transcription intended for specialist phonetic research will require different choices from one produced for a general audience or for lexicographic purposes. Clearly identifying your purpose at the outset helps you avoid unnecessary work and ensures that your transcription remains readable and useful for its intended users.

For example, at this stage, you should decide whether your transcription will be broad — recording only the most phonologically distinctive contrasts with minimal diacritics — or narrow — representing as much phonetic detail as possible. This choice should be aligned with your research objectives and explicitly documented.

Multiple transcription systems have been developed for the documentation of Italian dialects. Each reflects different theoretical assumptions and historical traditions of study. The choice of system must be made deliberately, as it conditions both the nature of the categories used and the interoperability of the resulting data.

Widely used among Italian dialectologists is the “Ascoli–Merlo” system. With some modifications, this system was employed in a major documentation project of the Italian dialect heritage known as the “Carta dei Dialetti Italiani”. Substantially different transcription systems were used in works of fundamental importance for Italian dialectology, namely the “Atlante Italo Svizzero” (AIS) and the “Atlante Linguistico Italiano” (ALI). To this plurality of systems, the IPA was added in the second half of the twentieth century, initially facing resistance among Italian dialectologists and only becoming widely established in recent decades. The "Rivista Italiana di Dialettologia" adopts the IPA system. However, its first issue also proposed a simplified transcription system based on Italian orthography.

The existence of multiple systems does not merely represent a variation in symbols: each embodies a different approach to the categorisation and classification of phonetic reality, and must be understood within the historical and theoretical context in which it was developed.

For any new transcription project, the use of the International Phonetic Alphabet (IPA) is strongly recommended. The IPA is the most widely used phonetic transcription system internationally, grounded in a modern and consolidated phonetic theory. It offers extensive documentation of its symbols and principles, remains relatively stable over time, and provides a flexible yet relatively simple system: the most phonologically relevant distinctions are represented through alphabetic symbols, while diacritics allow for finer phonetic detail. It is also supported by extensive reference materials, including the Illustrations of the IPA series.

If your project is already underway and uses a non-IPA or mixed transcription system, you are not required to convert it to IPA. However, you should provide clear and exhaustive documentation of all symbols employed and of the transcription practices adopted. In this context, the IPA can serve as a metalinguistic reference to clarify the phonetic values of non-standard symbols. If you decide to attempt a conversion to IPA, you should keep in mind that transcription systems are not strictly isomorphic, and that a one-to-one correspondence between symbols is not always possible. Any conversion should therefore preserve the original transcriptions, which users should always remain able to consult.

In some contexts — particularly for dialects with an existing written tradition — transcriptions seek a compromise between orthographic local spelling conventions and phonetic notation; in such cases, all non-standard symbols used must be explicitly listed and linked to their IPA equivalent to ensure interpretability for any reader.

An example of a mixed approach is the orthographic system used to render Neapolitan dialect dialogue in Italian cinema. Stromboli ("Il dialetto sul grande schermo. Il napoletano nella storia del cinema italiano", Firenze, Cesati, 2022) follows the written tradition of Neapolitan, itself based on Italian orthography, while adding a small number of phonetic symbols from different sources. The following extract is from a scene in Massimo Troisi's "Ricomincio da tre":

"E cchéllə ca dico iə, cu tutt' 'a ggènte ca čə sta a Nnapulə pròpio nu móstrə éva vənì 'e cas' affianc' a mme?'"

Here, the schwa symbol [ə] is used to represent the unstressed indistinct vowels typical of Neapolitan, and ‹č› stands for IPA [tʃ] (as in čə = 'there is' where čə = [tʃ]). These conventions must be explicitly glossed in any published transcription.

Whichever system you adopt, a solid command of its symbols and conventions is essential before beginning. Study the relevant charts and reference materials carefully, paying particular attention to the sounds present in the dialect you are documenting. If working with IPA — the recommended choice — the authoritative references are the "Handbook of the International Phonetic Association" (Cambridge UP, 1999) and the Illustrations series published in the Journal of the International Phonetic Association, which provides transcribed samples of many languages and dialect varieties. For a broader and more comprehensive understanding of the modern phonetic theory underlying the IPA system, you should also consult J. Laver ("Principles of Phonetics", Cambridge UP, 1994).

The traditional and methodologically recommended approach for dialect data is impressionistic transcription: the transcriber aims to represent, as faithfully as possible, their auditory impressions, mapping them onto the taxonomic categories provided by phonetic theory. This is neither a phonological idealisation nor an attempt to normalise pronunciation toward a reference model. Rather, the transcription records what is actually perceived as produced by the speaker in a given utterance at a specific moment.

Impressionistic transcription requires that the transcriber have a trained ear and a thorough command of the phonetic system being used. The availability of digital recordings and acoustic analysis software (see Tools section below) greatly facilitates this work by allowing the analyst to listen repeatedly to the sequence to transcribe and, where necessary, to complement auditory perception with the visual inspection of acoustic properties.

After the initial transcription, and having respected the phonetic reality produced by the speaker, it may be appropriate to intervene with simplification and normalisation operations to make the transcription usable for its intended purpose and audience. 

One common case concerns the normalisation of pronunciation variants. Minor fluctuations in the pronunciation of a given segment — which may have been transcribed with slightly different symbols across occurrences in the corpus — can often be unified under a single symbol after analysis, provided that the variation carries no contrastive or analytically significant function. For example, if a vowel is inconsistently realised between two slightly different articulatory positions, the analyst may choose either to preserve this difference by using two distinct phonetic notations, or to transcribe it uniformly using a single notation once it has been established that the variation is not meaningful to the analysis. The decision to normalise must follow, not precede, analysis: only after examining the distribution and function of a variable should you decide whether to represent its different variants with a single symbol. Normalisation should be conservative — when in doubt, preserve the variation.

A “narrow” transcription is appropriate for strictly phonetic research, or for studies in which phonetics constitutes a privileged level of analysis. Even in such cases, however, the transcriber may choose to represent certain phenomena in greater detail when they constitute the direct object of investigation, while paying less attention to other aspects. Although such choices may be acceptable for practical reasons related to time and resources, they must always be made explicit so that readers do not draw inaccurate conclusions from the transcription data. Any deviations from standard IPA symbols and conventions should likewise be clearly explained and properly justified, and should only be introduced when the IPA does not already provide a notation sufficiently adequate for what is being transcribed.

By contrast, in research that does not pursue specifically phonetic objectives, “broad” phonetic transcriptions are generally preferred. These omit most diacritics and tend to normalise minor fluctuations in pronunciation. More generally, every decision taken during transcription and post-processing should be documented clearly and transparently, so that users of the transcription can interpret the data correctly.

Prosodic annotation in dialect transcription is generally kept to a minimum. Vowel length is typically marked with [ː]; consonant length may alternatively be rendered by repeating the consonant symbol, a solution that many transcribers prefer for readability. Word stress is conventionally indicated with [ˈ] before the stressed syllable, omitted on weak monosyllables but retained on strong ones. For prosodic boundaries, [ | ] and [ ‖ ] are widely used to mark minor and major pauses respectively; silent pauses are often rendered with (.), following Conversation Analysis practice. For spoken-language phenomena not fully covered by IPA — such as hesitation lengthening or abrupt interruptions — additional symbols like ⟨ ⟩ and + are commonly adopted, on the condition that their values are explicitly declared in the conventions document. Words are generally separated by spaces, as this considerably aids the reading and comprehension of the transcription.

Once the transcription has been reviewed and finalised, it is ready to be shared with its intended audience — whether as part of a published article, a corpus deposit, or an open-access dataset.

Ultimately, a successful transcription process strikes an optimal balance between phonetic reliability and textual readability, ensuring both the accurate representation of dialectal speech data and its long-term scientific interoperability.

Giovanni Abete

This workflow outlines the main steps involved in the transcription of dialectal speech data, with particular attention to the Italian linguistic context and the Italo-Romance dialects. Although it has been primarily conceived for the Italian reality, the workflow may also be adapted and applied to other geographical and linguistic contexts, provided that the necessary methodological and linguistic adjustments are taken into account. This process is documented by Giovanni Abete in the chapter “Trascrivere il dialetto” (pp. 211–217) of the "Vademecum per il trattamento delle fonti orali" (Tavolo permanente per le fonti orali, 2023).

Edit actor

Other URLs relevant to the resource, e.g. ORCID, etc.

URL, ideally landing page, of the resource.

Person/group/body who helped create the resource.

Should be concise and raise interest in the entry. Allows <Link>Markdown</Link> content.

Other URLs relevant to the resource, e.g. code, DOI, etc.

Upload media

URL pointing to info about the access policy.

The activities you can do with the resource. See <Link></Link>.

Name of the conference where paper was given.

Describes the discipline covered by resource. See <Link></Link>.

Locations where the Resource is offered. See <Link></Link>.

URL to helpdesk for incidents & user requests.

Audience targeted by the resource. See <Link></Link>.

A particular published issue of a journal.

Concept or term related to MP entry. See <Link></Link>.

Language(s) in which a resource is available. See <Link></Link>.

Select license pertaining to resource. See <Link></Link>. If license is unknown, use the "Terms of use" property field to describe the usage terms.

Status of the Resource life-cycle. See <Link></Link>.

Mode of use for the resource. See <Link></Link>.

File format of the linked resource. See <Link></Link>.

Page #s of a resource, separate by commas.

Refine the resource's nature and function. See <Link></Link>.

Links to non-MP materials that are relevant.

Info about performance levels provider expected to deliver.

Standard used by the resource. See <Link></Link>.

Technology Readiness Level of the Resource. See <Link></Link>.

Resource user manual and documentation URL.

Version of the tool or service that is linked.

The volume of a journal or multivolume book.

Edit source

Name

If you are having trouble with the sign in process, please <ContactLink>contact the SSH Open Marketplace team</ContactLink>.

If you are having trouble with the sign up process, please <ContactLink>contact the SSH Open Marketplace team</ContactLink>.

How to phonetically transcribe dialects

Related items(3)

Vademecum per il trattamento delle fonti orali

Treating Oral Sources in SSH Research

Transcription de données de langage oral

Workflow steps(8)

1 (Preliminary phase) Obtain high-quality audio recordings

2 (Preliminary phase) Define your research objectives upfront

3 Choose a transcription system

4 Learn or review your transcription system

5 Transcribe faithfully

6 Post-process and normalise the transcription

7 Annotate prosody

8 Publish your transcription