How to spell Patois, and why the same word looks different every time

Jamaican Patois has no official spelling. It is spoken far more than it is written, no single system is in general public use, and most written Patois is improvised on the spot out of English spelling habits — which is why the same word reaches you three or four different ways in a single afternoon. Two systems are worth knowing: the popular English-based spelling that people actually write to each other, and Cassidy/JLU, a phonemic system in which one sound is always written the same way.

Why one word turns up spelled four ways

Around three million people speak Patois in Jamaica and several million more across the diaspora, and almost none of them were taught to write it. English is Jamaica's official language — of law, of government, of the classroom and of the examination paper. Patois is the language of the house, the shop, the studio and most of the actual conversation on the island. For most of its history it lived almost entirely in the mouth.

So when somebody needs to put it on paper, or into a message, or under a photograph, they reach for the only letter-to-sound habits they have, and those habits are English ones. The result is a writing practice with no committee behind it: conventions that different writers hold to different degrees, drifting between families, between parishes, between one generation and the next.

That is a normal condition for a language rather than a deficiency in this one. English spent several centuries in the same state and was pulled out of it slowly, by government clerks, by printers and eventually by dictionaries. Nothing has applied comparable pressure to Patois, because the places it lives do not reward standardisation the way a printed book does.

What the everyday spelling is good at

It is easy to look at four spellings of one word and conclude that the system is failing. It is succeeding at a job that is simply not the job a learner wants done.

The popular spelling is optimised for instant recognition by somebody who already reads English and already speaks Patois. You spell the word so that the reader's eye takes it in without stopping: wata, mawnin, likkle— little, pickney— child, nuff— plenty. Each uses English letter-to-sound habits, so it is recognised on sight, and each sits far enough from an English word to signal that you are reading Patois.

Judged on that measure it is a good system, and it has one property no invented system can buy: it is what the reader is already used to. A Jamaican reading likkle in a message has met that word-shape countless times. Change the spelling to something more principled and you have made the sentence slower to read for everyone who reads it fastest today. That is also why our own pages lead with it.

The one thing it cannot tell you

What the popular spelling cannot do is tell somebody who does not already know the word how it is pronounced. English letter-to-sound habits are ambiguous even for English. Applied to a different sound system they are worse than ambiguous: they hand you an English pronunciation you already have, with no signal that it is not the one being asked for.

The problem is structural rather than careless. A writing system built out of another language's spelling conventions can only distinguish the sounds that other language distinguishes, so every sound Patois has and English does not falls straight through the gap. Three of those gaps do most of the damage, and a learner who knows them reads popular spelling far more accurately than one who does not.

Three places the English spelling misleads

The word that looks like an English tree

Patois shifts the English cluster in tree, true, train and trouble. Where English has a plain t before the r, Patois has an affricate — the sound English writes as ch — so the cluster is chr rather than tr. The number between two and four is therefore not said the way the popular spelling tree suggests. The phonemic form is tree(chrii)— three, and the same shift runs through true(chruu) and the rest of that family.

This is the most consequential row in Patois spelling: the popular form is identical to an English word pronounced differently and meaning something else, and nothing on the page tells you there is anything to notice.

The same shift, voiced

Do the same thing to d followed by r and you get the voiced partner: the sound English writes as j. So drink is drink(jringk) and dry is dry(jrai). Again the popular spelling is the English word unchanged, and again it hands the reader an English pronunciation with no warning attached.

Vowels that English writes badly

English vowel spelling is the least reliable part of English spelling, so it is the least reliable thing to borrow. Take the Patois word for day. It is written day, which tells an English reader to say the English word. The vowel is a different one, a diphthong the phonemic system writes ie, giving day(die). That form is alarming to look at, which is exactly why nobody writes it in a message — and exactly why a learner needs to have seen it once.

The same diphthong shows up in name(niem), and its back-vowel partner in go(guo) and stone(stuon). Vowel length does real work too: seat(siit) and sit differ only in how long the vowel is held, and popular spelling marks that difference only when English happened to mark it first.

Doubled letters that mark nothing

Look at likkle, pickney, dutty, duppy, tallawah and nuff and you will notice a habit: doubled consonants everywhere. In English a doubled consonant carries information — it usually says something about the vowel in front of it, which is the entire difference between hoping and hopping.

In popular Patois spelling it carries none. The doubling is a word-shape convention: it marks the word as Patois rather than English, it looks right to a reader who has seen it before, and it says nothing whatever about length, stress or vowel quality. The phonemic forms are likl, pikni, doti, dopi, talawa and nof, each shorter than the popular spelling with nothing lost.

That does not make the doubled spellings a mistake. They are settled shapes and they are what people read. The point is narrower: do not read information into them. The doubled k in likkle is not telling you anything about the vowel in front of it, and the letters after it are not a vowel either — the word ends in a syllabic l, which is why the phonemic form is likl and stops there. The decorative final h in tallawah belongs to the same family. It is a flourish, not a sound.

The apostrophe habit, and what it implies

The habit worth breaking is the apostrophe: da', 'dem, wid', gwan'. The logic is obvious enough — the writer knows the English word has a letter there, hears that the Patois word does not, and marks the absence the way English marks a contraction.

It is worth saying plainly what that punctuation asserts. An apostrophe means something has been removed here. Using one in Patois says that the English word is the underlying form and the Patois word is that form with pieces knocked off it. That is a claim about the language, made in punctuation, several hundred times a page.

The claim is not true. respec(rispek) is not the English word with its final consonant deleted; it is a word in a language that does not put that cluster at the end of a word. dem— they, them is not them with a letter missing — it is a pronoun that also does the work English does with a plural ending. These are the forms, not damaged copies of other forms, and the question underneath is worked through in full in our guide to whether Patois is a language or a dialect.

Words with no English spelling to borrow

Improvised English spelling has one more limit, and it produces the widest scatter of all. It works by finding the nearest English word and adjusting it, so when there is no nearest English word it has nothing to hold on to.

Part of the Patois vocabulary came from West and Central African languages, carried across during the centuries of the slave trade and kept. Those words have no English cognate to spell them against, so each writer transcribes from scratch and different writers land in different places. nyam(nyam)— to eat is stable enough — it is short, and the everyday spelling and the phonemic one happen to land on the same letters. The second-person plural pronoun is not: you will see it written unnu, unno, oonoo and unu, depending entirely on who is typing.

The phonemic system picks one and holds to it — unu, each vowel written once, no doubling left to decide about. It cannot settle which words came from where, which is history rather than spelling, and we take it up in our guide to the Patois words that came from West African languages.

What a phonemic system fixes

A phonemic writing system makes one promise and keeps it in both directions: one sound is always written the same way, and one written form always says the same sound. Everything else follows from that promise.

You can read a word you have never seen and pronounce it. You can hear a word you have never seen written and spell it. Neither is possible in English, and neither is possible in popular Patois spelling. A machine can do the same thing, which is why every phrase on this site carries a phonemic form underneath the one you read, and why a speaking Exercise can line your voice up against the sounds of a phrase rather than against a guess.

It writes the things popular spelling drops. Vowel length is written, so siit and sit are visibly two words. Nasal quality is written, which matters in a language where kyan(kyan)— can and cyaahn(kyaahn)— cannot are opposites separated by vowel length and a nasal mark. The palatal glide is written, so gyal(gyal) and car(kyar) show a sound English spelling has no settled way to write.

The cost is the one you already saw in die for day. A phonemic system is unreadable on sight to somebody who has spent a lifetime reading English, because it deliberately refuses English's conventions. That is the trade it makes on purpose, and it is the whole reason a serious Patois resource carries two spellings rather than one.

How the system actually works — the vowels, the long vowels, the four diphthongs, the nasal mark, the consonants and the letters it does not use — is set out in our guide to the Cassidy/JLU writing system.

So which spelling should you use?

Write to your reader. Sending a message to your aunt, write it the way your aunt writes it: popular spelling, no apostrophes, whatever word-shapes your family already uses. Nobody is served by a phonemic transcription in a group chat, and a spelling that slows your reader down has failed at the only thing writing is for.

Reach for the phonemic form when the question is pronunciation — when you want to know how a word sounds, or to record how it sounds so somebody else can get it right. On this site you get both at once: the spelling you would write, and the spelling that tells you what to say.

When you meet a Patois word spelled a way you have not seen before, the useful reflex is not to wonder which spelling is the right one. It is to say the word out loud. The spelling is a suggestion; the language is the sound. The ordered route through the grammar starts at our guide to how to speak Jamaican Patois.