Writing A Language Down For The First Time

Giving a language its first written form looks like a technical task and is mostly a political one. Every choice made at the start becomes very hard to change afterwards.

Deciding which sounds count

Speech contains far more distinctions than a writing system records. The first job is working out which differences change meaning and which are simply variation between speakers.

Only the meaning-changing ones need symbols. Recording everything produces a system too fine to write quickly, and recording too little produces one that cannot distinguish common words.

Tone and length are the usual difficulties. Both carry meaning in many languages and both are easy for an outside analyst to under-record, because they are unfamiliar categories.

Choosing a script has consequences

A new orthography almost always borrows an existing script, and the choice aligns the language with a neighbour, a religion, a colonial history or a state.

The borrowed script rarely fits. Sounds with no symbol have to be handled by adding marks, combining letters or repurposing letters the language does not need.

Each solution has costs. Extra marks are slow to write and easy to omit, letter combinations lengthen everything, and repurposed letters confuse anyone literate in the source script.

Which dialect becomes the written one

Writing forces a choice among varieties that were previously equal. Whichever variety the system is built around becomes the prestige form, and the others become deviations from it.

The choice usually follows population, administrative convenience or where the people doing the work happened to be based, none of which is a linguistic argument.

Speakers of the other varieties then face a permanent mismatch between how they speak and how they must write, which is a well-known source of difficulty in schooling.

Word boundaries are not obvious

Where one word ends and the next begins is far less clear in speech than writing makes it appear, and languages that attach many elements to a stem make it harder still.

Decisions about spacing look trivial and affect everything downstream, including dictionaries, indexing, searching and how children are taught to read.

Different projects working on related languages have made opposite decisions, which leaves neighbouring communities with written forms that look far more different than the speech does.

Why revisions are so difficult

An orthography becomes entrenched as soon as material exists in it. Books, signage, schooling and anyone already literate all constitute a constituency for the current system.

Reform proposals therefore arrive with a cost attached, and the people who benefit most are those not yet literate, who have the least influence over the decision.

Systems that were designed carefully with speakers in the room fare better than ones imposed and then corrected, which is the main lesson from a century of these projects.