The Word They Did Not Use

Most translation criticism has an original and a translation, and asks what the crossing cost. This one has neither. At Waitangi there is no original: there are two texts, both signed, and since 1975 New Zealand law has carried both. What they say is not the same, and the difference is not a matter of register. It is a difference about what was given away.

None of that is new, and none of it is disputed by anyone who has read the two texts side by side. What this page adds is not an argument but an instrument. The four documents are loaded and indexed word by word, so that instead of being told a word is missing you can type it in and watch the counter return zero. The sheets are photographs. Where our reading differs from a published transcription, we say which, and show you the line, and you can decide.

Choose a document and a passage. Underlined phrases are keyed: click one and its counterpart lights up in the other column, with a note on what moved. With JavaScript off, everything below is simply present in full.

All four texts above are indexed. Type a word and the counts are computed live, from the very paragraphs you have just read, over whole words only, so rangatira does not count rangatiratanga.

Both documents are officially bilingual, so each one tells you what its own drafters thought a word was worth. James Busby, the British Resident, composed the 1835 Declaration in English and the missionaries put it into Māori; Ruth Ross, who went looking, could not establish which of them did it. In 1840 Hobson brought an English draft to Henry Williams at about four in the afternoon of 4 February, and by nine the next morning Williams's translation was on Busby's table being gone over. Those two timestamps are what the record gives. How the hours between were spent is not documented, and the often-repeated “translated overnight by Henry Williams and his son Edward” rests on a family biography published in 1874. Set the two documents side by side and three of the words move.

You do not have to take the transcription on trust. These are the sheets. Contrast has been raised a little so that iron-gall ink reads on a screen; nothing else has been changed. On a narrow screen, scroll each plate sideways.

There is one more witness, and it has the advantage of being entirely uninterested in the question. Dictionaries of Māori were being compiled by the same missionaries, in the same decades, for their own purposes. What they contain, and what they do not, was settled long before anyone thought the wording of the Treaty was worth arguing about.

The obvious move is to look up kawanatanga. The more telling move is to look up the English.

So the missionary lexicon of 1852, twelve years after a Governor arrived and while one was still governing, offers no Māori for “governor” and none for “sovereignty.” That is not a claim about what the language could express. It is a fact about what the book contains, and you can see it.

What the same book does offer, in the same English-to-Māori half, is this.

Read those together and the shape is hard to miss. The word the missionary dictionary gives for kingdom and for dominion is the word the Treaty uses for what the chiefs keep. The word it gives for authority and for power is the word that appears in the Treaty not at all. And the word the Treaty uses for what the chiefs give up is not in the dictionary in either direction.

It is not in the earliest one either. Kendall and Lee's Grammar and Vocabulary of the Language of New Zealand, London 1820, has no kawana: the run at that point is Ka wádu, Káware, Ka wa téa, then over the page to Kaweka. It has no rangatiratanga, only Ránga tíra ; A gentleman or lady. And its only mana is the pronoun: Mána ; For him. The noun is not there.

The first printed dictionary entry for the word, as far as we could find one, is in Ernst Dieffenbach's word list of 1843, and it is tagged:

It is natural to imagine the two texts as equals: an English document and its Māori translation, one of each. The record is lopsided, and the lopsidedness is stranger than the lexical difference.

On the Māori side there is one text. It was read out on 5 February, debated, signed on 6 February, and then carried round the country on nine sheets, eight of them in Māori. More than five hundred signatures were gathered. Thirty-nine of them are on the English sheet, which was signed at Waikato Heads and at Manukau in March and April, not at Waitangi.

On the English side, the document that matters is missing. The English text printed in Schedule 1 of the Treaty of Waitangi Act 1975, and quoted whenever the Treaty's English is quoted, is not the paper Henry Williams translated from. Phil Parkinson's 2004 study of the drafts concludes it is a pre-translation draft with the date altered from the fifth of February to the sixth. The draft actually handed to Williams has been lost since at least 1869, when the Legislative Council formally asked for it and the Colonial Secretary's office replied that “The original draft (if any) is not on record.” It survives only through three transcripts made in 1840, all dated 5 February, whose preamble begins nothing like the sheet's: not “regarding with Her Royal Favor the Native Chiefs,” but “regarding with deep solicitude the present State of New Zealand.”

Parkinson puts the consequence bluntly, and it is worth having in his words: “There is no authenticated ‘translation of the Māori text’ from Henry Williams into English at all.” Hobson did send Gipps an English text certified by Williams as “as literal a translation of the Treaty of Waitangi as the idiom of the language will admit of.” It is not a translation of the Māori. It is another English draft.

So the shape of the thing is this. The text that hundreds of people signed is a single stable document in Māori. The text whose words are quoted when sovereignty is discussed is in English, exists in several states that differ from one another, and descends from a draft nobody has seen for over a century and a half. New Zealand's Parliament has said as much, in the preamble to the 1975 Act that gave the Treaty legal effect:

And there is one more turn. In 1869 the Legislative Council, having failed to find the lost draft, ordered a careful translation of the Māori text back into English. T. E. Young of the Native Department did it. Where the English text says the chiefs cede “all the rights and powers of Sovereignty,” Young rendered the Māori as nga tikanga me nga mana katoa o te Rangatiratanga. The government's own translator, twenty-nine years on, reached for rangatiratanga to name the thing that had been given away, which is the same word the Treaty uses for the thing that was kept.

Everything above is about words on paper. That is a narrow thing to be sure of, and it is worth being exact about how narrow.

A treaty is not only its text. The debate at Waitangi ran for a day, out loud, in Māori. Rangatira questioned Hobson and the missionaries and each other, and what was said in those speeches is part of the transaction and is not recoverable from the sheet. It is also the ground on which both sides of the modern argument stand: the Crown's case to the Waitangi Tribunal was that the oral explanations conveyed sovereignty, and the Tribunal's finding rests on the oral proceedings too, and goes the other way.

The best-known moment in that record is William Colenso's. Colenso had printed the treaty, knew the language, and was standing there when the first chief came forward to sign. He published his account in 1890, from notes; it is public domain, and it is worth reading in full rather than in summary, because the famous line is not the strongest one in it.

A word is not a dictionary entry. Showing that Kawanatanga renders “any function of Government” in 1835 is strong evidence about what the translators were doing. It is weaker evidence about what any particular rangatira heard. And the case cuts both ways: Pat Hohepa, a Māori-language authority, pointed out to the Tribunal that te Tiriti is about twenty sentences and under four hundred words, of which only thirteen are transliterations from English, and that such a document “would hardly be a matter that needed two days of intensive wananga to comprehend.” Against that, Waiohau Te Haara counted kāwana in about a hundred and sixty verses of the Māori Bible, the book those chiefs read most, and found it almost always named an office below a king, for which the usual word was rangatira.

The serious counter-argument is not the one you would guess. The strongest recent case for reconciling the texts, Ned Fletcher's The English Text of the Treaty of Waitangi (2022), does not argue that the chiefs ceded sovereignty and knew it. It argues that the English meant less than modern readers assume: that British officials in 1840 were not seeking to displace tribal government, so “sovereignty” in the English draft is narrower than it sounds and is reconcilable with tino rangatiratanga. Note where that leaves the two texts. It does not close the gap by making the Māori mean more.

And this page takes no position on the constitutional question. Whether sovereignty was ceded is live law and live politics in New Zealand, and no word count decides it. What the record contains is this. In 2014 the Waitangi Tribunal, reporting on stage one of its Te Paparahi o Te Raki inquiry, concluded that “the rangatira who signed te Tiriti o Waitangi in February 1840 did not cede their sovereignty to Britain,” and that they agreed instead to share power with the Governor. In the same report it said, in terms that are almost always dropped when the finding is quoted: “In drawing this conclusion, we say nothing about how and when the Crown acquired the sovereignty that it exercises today.” The finding covers the Bay of Islands and Hokianga signatories in February 1840, not all five hundred across nine sheets and seven months. On the day it was released the responsible minister said “There is no question that the Crown has sovereignty in New Zealand. This report doesn't change that fact.” Read carefully, those two statements do not contradict each other, and that is the single most often collapsed point in the whole dispute.

The Tribunal is a permanent commission of inquiry. Under the 1975 Act it has exclusive authority to determine the meaning of the two texts for the purposes of that Act; it recommends rather than orders, is barred from touching private land, and can bind only over certain categories of former Crown land. Its historical findings are authoritative within its remit and highly influential. They are not self-executing, and they are not a court's ruling on the law.

The textual argument itself is older than the Tribunal. It reached its modern form in Ruth Ross's 1972 article “Te Tiriti o Waitangi: Texts and Translations,” which is where the 1835 comparison, the missing English draft and the Young translation were all set out together, and which ends by calling the treaty “hastily and inexpertly drawn up, ambiguous and contradictory in content, chaotic in its execution.” Crown counsel at the Tribunal called her analysis “superficial.” Both of those are on the record, and we have printed the documents so that you can see what is being argued about.

This is the venue's first entry of a kind we have not built before: not an original and a translation, but two texts of equal standing that disagree. It is worth naming one other, because the comparison shows how unusual the New Zealand outcome is.

On 2 May 1889 Italy and Ethiopia signed the Treaty of Wuchale in Italian and Amharic. Article 17, in Italian, says the Emperor consente di servirsi of the Italian government for his dealings with other powers: he consents to make use of it. The Amharic verb is ይቻላቸዋል, which is not a permission granted but an impersonal possibility: it is possible for him to do so. One text creates an obligation; the other records an option. Italy notified the powers that Ethiopia was its protectorate. Menelik repudiated it, and in 1896 an Italian army was destroyed at Adwa.

What makes Wuchale the sharpest case is Article 19, which warrants that the treaty was drawn up in both languages and that “the two versions agreeing perfectly with each other,” both texts are official and have equal authority. A clause guaranteeing that the two texts agree, in a treaty whose two texts do not. When peace was finally made in 1896, the treaty was drawn up in Amharic and French. Italian was left out.

Set beside that, the New Zealand response is genuinely unusual. Other empires resolved the problem in advance by fiat: the 1858 Treaty of Tientsin simply provides that where the English and Chinese differ, the English shall be held correct. Wuchale resolved it by war. New Zealand did something else: in 1975 Parliament wrote the disagreement into the preamble of a statute, printed both texts in a schedule, and created a body to adjudicate between them. Whatever one thinks of how that has gone, it is a rarer answer than either of the others.