ToBI
ToBI (which stands for Tones and Break Indices) is a set of rules for writing down how speech sounds. It is used to show the melody and rhythm of sentences.[1][2]
Scientists, including Mary Beckman and Janet Pierrehumbert, created ToBI in the 1990s. They wanted everyone who studies language to have one shared way to write down speech sounds. This helps with language research and with making computers that can talk and understand speech.[3]
Instead of just saying "the voice goes up," the ToBI system uses special labels for different parts of the speech melody. These are called pitch accents, phrase accents, and boundary tones. It also has a way to mark the size of the pause or break between words. These are called break indices.
ToBI was first made for American English, but it has been changed to work for many other languages.
How ToBI was made
The main idea for ToBI came from a theory by Janet Pierrehumbert in 1980.[4] This theory said that the melody of speech is made of simple high (H) and low (L) tones. These tones are linked to certain parts of words, like the stressed syllable or the end of a sentence.
In the 1990s, scientists needed one clear system for writing down speech sounds. They were building large collections of recorded speech to train computers to speak (text-to-speech) and understand speech (speech recognition). ToBI was made to be a clear and trusted system that all scientists could use in the same way.[5]
The main parts of ToBI
A ToBI description is written on different lines, called "tiers." The most important tiers are the tone tier and the break tier.
The tone tier
The tone tier shows the melody of the speech using three kinds of tones:
Pitch accents
Pitch accents are tones put on a stressed syllable in a word. This makes the word stand out. They are the basic building blocks of speech melody. The most common pitch accents in English are:
- H*: A high sound on the stressed part of a word. This is often used to show what is important.
- Example: Marianna ate the noodles. (The sound is highest on noo-).
- L*: A low sound on the stressed part of a word. This can make a sentence sound finished.
- Example: It's an absolute fact.
- L*+H: The sound starts low and then goes up quickly.
- Example: I can't believe it!
- L+H*: The sound rises to a high point on the stressed word. This is often used to show a choice between two things.
- Example: I said her mother, not his mother.
- !H*: A high sound that is lower than the high sound that came before it. If you have many of these in a row, the sound goes down like steps.
- Example: It's an absolutely fantastic novel.
Phrase accents
A phrase accent is a tone that controls the sound between the last stressed word and the end of a group of words.
- L-: A low phrase accent.
- H-: A high phrase accent.
Boundary tones
A boundary tone is a tone at the very end of a sentence or thought. It shows if the voice goes up or down at the end.
- L%: A low boundary tone. The sound stays low at the end.
- H%: A high boundary tone. The sound goes up at the end.
Phrase accents and boundary tones work together to make the final sound of a sentence.
- L-L%: The most common sound for a simple statement in English.
- H-H%: A "rising" sound, used when the speaker plans to say more.
- L-H%: A common sound for yes/no questions.
The break index tier
This line uses numbers to show how much of a pause or break there is between words.
- 0: No break at all (like in gonna for going to).
- 1: A normal space between words. This is the most common break.
- 2: A very small pause between words, smaller than a comma.
- 3: A short pause. This marks the end of a small group of words (an intermediate phrase).
- 4: The biggest pause. This marks the end of a full sentence or thought (an intonational phrase).
Example of ToBI writing
Here is an example of how ToBI is used to write down a sentence.[6]
| Word | Tonal Events | Break Index |
|---|---|---|
| Marianna | H* | 1 |
| made | (none) | 1 |
| the | (none) | 1 |
| marmalade. | H* L-L% | 4 |
This writing shows:
- A high sound (H*) on the stressed parts of Marianna and marmalade.
- The sound goes down at the very end (L-L%), which is normal for a statement.
- There are normal breaks (1) between the words.
- There is a big break (4) at the end of the sentence.
Why ToBI is important
ToBI is now the main system that scientists use to study the melody of speech. It is used for:
- Language Research: To study and compare the sounds of different languages.
- Computer Voices: To help make computer voices (text-to-speech) sound more real.
- Speech Recognition: To help computers understand human speech better.
- Brain Science: To study how people use speech sounds to understand each other.
ToBI for other languages
The ToBI system can be changed to work for many languages. These new versions often add a letter to the name, such as:
- Gla-ToBI for Glasgow English
- J-ToBI for Japanese
- G-ToBI for German
- K-ToBI for Korean
- Cat-ToBI for Catalan
- Sp_ToBI for Spanish
- P-ToBI for Portuguese[1]
- Sonsorol-Tobi for Sonsorolese, spoken in regions of Palau and Micronesia[7]
Related pages
References
- ↑ 1.0 1.1 "Portuguese ToBI | about ToBI". labfon.letras.ulisboa.pt. Retrieved 2025-07-23.
- ↑ Jun, Sun-Ah (2022). The ToBI Transcription System: Conventions, Strengths, and Challenges (PDF). MIT Press.
- ↑ Silverman, K., Beckman, M., Pierrehumbert, J., Ostendorf, M., Wightman, C., Price, P., & Hirschberg, J. (1992). ToBI: A standard for labeling English prosody. In Proceedings of the 1992 International Conference on Spoken Language Processing (ICSLP), Vol. 2, pp. 867-870.
- ↑ Pierrehumbert, Janet B. (1980). The Phonology and Phonetics of English Intonation. PhD dissertation, MIT.
- ↑ Beckman, M. E., & Ayers, G. M. (1997). "Guidelines for ToBI transcription". In The Ohio State University Research Foundation. Available online.
- ↑ Administration. "Intonation - ToBI introduction". Macquarie University. Retrieved 2025-07-23.
- ↑ "Grammar and vocabulary of the language of Sonsorol-Tobi". dokumen.pub. Retrieved 2025-07-23.