Automated Syllabic Conversion

Cree Syllabic Transliteration on the Desktop

The debate about the best written form for the Cree language seems destined to continue for some time in Western Canada. Some communities prefer “Standard Roman Orthography” (SRO) which uses alphabetic letters (just like English); other communities prefer Syllabics.

Examples of published Cree are rare in either writing system, but to support literacy efforts and reach the greatest number of Cree readers, it seems practical for new materials to appear in both. Hand-conversion of SRO to syllabics, or “transliteration”, is painstaking work that requires enormous concentration and a great deal of proofreading. It also demands biliteracy and a solid understanding of the conventions of both systems. Since much of the task is repetitive, though, computers can be used to streamline the process significantly.

In the mid-80s, John D. Nichols, then at the University of Manitoba, developed a method for conversion of alphabetically-written Ojibwe to syllabics. His first design used a mainframe computer to process the texts, which were then downloaded and printed using a custom font on a dot matrix printer.

With the arrival of PCs and desktop laser printing, it was necessary to adapt the conversion algorithms, and design new laser fonts. This work fell to me, under the direction of Nichols. With a few additional changes, the conversion became applicable to Cree.

The greatest advantage of computer-automated transliteration is that errors introduced by re-writing by hand, then re-typing can be avoided. If the SRO input is correct, the syllabics will reflect that accuracy. Proofreading is essential, as always: automated conversion helps eliminate “new” errors. In fact, conversion sometimes even makes “old” errors easier to spot. The process is a good illustration of the old computer maxim: Garbage In, Garbage Out. Minor inaccuracies in the SRO can turn into glaring errors in the syllabic output.

The conversion algorithm — which remains the same regardless of the platform or the software that is being used — is essentially a list of strictly ordered rules, in which each possible syllable (e.g., pwaa) is replaced with the corresponding Syllabic symbol.

Examples of syllabic texts created through this process, and printed in the custom-designed Syllaco font, appear in a number of volumes published by the University of Manitoba Press and elsewhere. I am honoured to have a sample of this work included Robert Bringhurst’s 2008 Surface of Meaning: Books and Book Design in Canada.

I was invited in 1987 to demonstrate and teach this conversion method at the Saskatchewan Indian Languages Institute, and prepared an instruction manual at that time. In 2008, with the further evolution of computers, and the introduction of Unicode fonts, which include standard definitions for syllabic characters, time, money, and interest may further adapt the process to contemporary technology.

 

Leave a Reply

Your email address will not be published. Required fields are marked *