Indic Fingerspelling Project

Ref: IF-PUB1/0a

Project URL: http://www.altruists.org/projects/if/

Prepared by: Dr. R. J. G. Upton & Miss S. H. Cho

Contact: if@altruists.org

Last Modified: 2003-01-31

1. Fingerspelling What is Fingerspelling?

2. How Many Hands? Should this be a 2-hand or 1-hand system?

3. Linguistic Fundamentals Signing conjuncts, different vowel forms etc.

4. Necessary Compromises Ease of use, clarity & inter-language issues

5. How to Help The Project Taking the project to the next stage...

1. Fingerspelling

Sign language is a system of communication where the units of communication are not words but gestures, including hand gestures, facial expressions and body movements. Fingerspelling, sometimes referred to as a Manual Alphabet is an extension of this system to complement sign language communication between individuals.

As the name suggests, it involves a series of hand gestures that spell out a word letter by letter. It is used most commonly either to translate from a written language in cases where no sign exists for a particular word (e.g. for names of people or places) or else where the sign for a particular word is unknown or has not yet been standardized between users.

Whereas the English alphabet has only 26 letters, Indic languages have about 50. The major Indic languages can be represented by a joint alphabet of about 65 characters (and innumerable more conjunct characters). This is not the only important different when designing an Indic finger system; the writing systems are much more complicated than the English, and so the one-to-one correspondence between letters and signs is not as straightforward.

2. How Many Hands?

1-handed systems are easier to sign but harder to interpret; they are used for speed such as in a one-to-one context or between familiar sign language users.
2-handed systems allow a wider variety of gestures; they are slower to sign, but easier to interpret, so are used where clarity is important, such as for addressing a large audience.

The number of letters in the Indic languages means that the system must be capable of a much wider variety than most other such systems. A system of 50 single hand gestures would require much more complicated gestures and so some of the speed advantage would be lost. A 2-handed system is more suitable for Indic Fingerspelling because it allows for a much wider range of gestures and hence greater clarity. Indian words tend to have fewer characters than their English equivalents, so although individual gestures may take longer (just as Indian letters take longer to write than Roman), the speed issue would be offset somewhat.

3. Linguistic Fundamentals

There are a number of decisions that need to be made before commencing work on a Fingerspelling system. Although these are interconnected, we will list them separately below:

#	Topic	Notes	Ambiguities	Issues
1	Inherent Vowel	Shoray-O is pronounced but not written in Bangla text.	Word pronunciation	Should these be signed or not?
2	Irregular vowel forms	Some vowels are written differently according to which consonant they follow.		Is it important to reflect this in the sign system?
3	Vowels	Vowels are written differently according to whether or not the follow a consonant.		Is it important to reflect this in the sign system?
4	Conjuncts	When consonants occur next to one another in a sentence, their written form often changes to show that there is no inherent vowel in between.		How do we sign conjuncts – is it enough just to sign the component letters?

A sign for the inherent vowel (and one for halant – to signify the absence of an inherent vowel) would provide a mechanism for conveying pronunciation and would solve the ambiguities surrounding vowels (3) cited above. Although important to have it for exceptional cases, the expectation is that in general use (e.g. for deaf & dumb) it would be rarely used, since omitting it would allow users to increase signing speed by about 20%.

Irregular vowel forms (2) are an ambiguity of writing, not of meaning, and are different for each Indic language. The system aims to transmit meaning, not necessarily writing, so it should not make different signs in this case.

Conjuncts (4) occur relatively infrequently and it is usually not important to treat them separately, as it is rare that this could lead to a miscommunication. Again, however, the system should contain a mechanism for purposes such as dictation. We propose to represent conjuncts not by introducing a separate gesture for them but by establishing a standard as regards the gesturing process. When signing a conjunct character, the user makes the gesture just below the previous gesture to indicate that it is part of the same character.

4. Compromises

The human hand can only make so many different gestures, and relating them to any written alphabet is more an art than a science, but there are plenty of general principles. The design of any Finger Spelling system involves managing the following trade-offs:

The easiest signs should be reserved for letters that occur most frequently, while more laborious ones are allocated to less common letters. Frequent letters should also be the clearest, but the clarity of a gesture cannot be defined in isolation; a sign is only as clear as it is unambiguous – a gesture is clear if the alphabet contains no others that look similar to it.

In the case of a multi-language system another level of compromises has to be made, which is worth illustrative with a pictorial example. Suppose we are devising 5 gestures for two languages, each of which has 2 letters of its own and 1 in common with the other language. It should be intuitively clear that the most distinctive gesture of the 5 should be used for the letter that is common to both languages. The situation could be depicted as shown on the right, using symbols to represent gestures:

What should also be clear by now is that the clarity within both languages could be greatly improved by swapping a pair of the signs around, as shown below.

The fundamental idea of the multi-language system remains that each letter should have a unique gesture. The observation that certain gestures are unlikely to ever be used in the same context does not compromise this – it merely allows for greater clarity within each individual language.

Letter frequency data is important to determine which ones occur more frequently and should therefore be given which signs. Such data for Bengali shows a very wide spread of frequency (from 14% down to below 0.01%) and we are looking to collect if for the other major Indic languages. Although a considerable degree of similarity is to be expected across languages, it will no doubt vary to a certain extent and so some compromises will have to be made.

Some compromises will have to be made as regards clarity and ease of signing but a lot can be done to minimize their effect. Moreover, the inter-language operability of a pan-language approach will produce a more convenient system that is a much better aid to communication than any sets of individual language-specific systems.

5. How to Help The Project

We are looking for volunteers with a heart for the project and with the following skills:

Pilot Project Coordination - Drawing up guidelines and feedback forms for institutions to participating in the pilot project

Publicity & Communications - Liaison with the appropriate authorities and preparation of press releases and other publicity material (web development skills would be an advantage).

Information Technology - The system will be distributed over WWW by incorporation into a piece of training software. We are also planning to make and distribute a free Indic Sign Language font.