That Mandan is not Welsh

There is a common myth that the Mandan people of North Dakota are in fact the decendants of Welsh colonists who sailed form Wales under Prince Madog. Do a web-search for “Mandan” with “Welsh” and you will find a plethora of pages promoting this story. Various pieces of evidence have led to this claim, most notably: Fair-skin/blue-eyes, Welsh style boats, and of course, that Welsh and Mandan speakers understand each other.

There have been word lists published that apparently show evidence that Welsh and Mandan are one and the same language (or at least Mandan is a product of Welsh). I would like to use this space to claim that these theories are patently incorrect. Welsh is an Indo-European language, Mandan belongs to the Siouan family.

The two language’s sound systems couldn’t be more different. Welsh has about 21 consonant sounds, Mandan has at most 11. Welsh has a lavish vowel inventory numbering 23 (including diphthongs), Mandan has 9 (including 3 nasal vowels not found in Welsh). Welsh uses 4 liquids: /l/, /ll/, /r/, /rh/, but Mandan has only /r/. I could go on and on, e.g. Mandan has a glottal stop, Welsh does not.

Mandan grammar is charactaristically North American, with long sequences of prefixes and suffixes on the verb, while Welsh is very European, marking tense and subject agreement only using portmanteaux suffixes. Welsh, like other Celtic languages, is famous for its system of mutations (changing the sound of the first letter of a word in certain lexical and grammatical environments). Mandan certainly does not do this, but instead, rotates between /s/, /š/, and /x/ for semantic reasons (e.g. /š/ being more intense than /x/, as in ‘pušak’ – ‘course grained’ vs puxak – ‘fine grained’).

I hope the vocabulary examples below magnify the differences.

Vocabulary

Numbers

Examples of Grammar

Of all these examples, the two languages share only one word in common, ‘house’. Statistically, there is bound to be at least one or two common words between languages, this does not imply they are related. That the word for ‘many’ in Korean is [mani], and ‘two’ is [tu(l)] does not imply a Korean colonisation of England in the distant past. Nor does the fact that the Mohawk pronominal affixes for 1st and 2nd person—/k/ and /s/— match the Hungarian affixes make it reasonable that the early Iroquois settled Hungary. These sorts of similarities are bound to happen. Historically, Mandan ‘ti’ would be related to /tʰipi/ in Dakota, while Welsh ‘tŷ’ is derived from Indo-European *tegos meaning cover/roof (English thatch). So these two words, although superficially the same, stem from different sources.

To be fair, I will present some data which people have used to show Mandan to be Welsh. The information below is from a letter by George Catlin (in North American Indians, Peter Matthiessen ed., Penguin 1989), an early American explorer of the interior of North America. Along with cultural evidence that I won’t speak about, he gives the following word list. The modern Welsh and Mandan words are given in [square brackets], proper translations are given in (parentheses).

As can be seen, the correct Mandan words (where I have been able to find them) are somewhat similar to what Catlin recorded – assuming a sound change from ‘r’ to ‘n’. However, Mandan uses pronouns (the majority of the list) vanishingly rarely. These are not in fact pronouns, but pronominal prefixes, which almost always have to be placed at the beginning of a verb root, so that a better translation of English ‘I’ or Welsh ‘fi’ is ‘wįoʔra’ – ‘it is me, that being me’. By taking a Mandan pronominal prefix and identifying it with a Welsh pronoun—which occurs after the verb, Catlin made a huge mistake in judgement. In any case, Mandan is what is called an active-stative language (common to several North American Native language families). This system of organising nouns in relation to the verb is quite different from the nominative-accusative model of Welsh.

What is a sentence in Welsh or English is a single word in Mandan:

• wa:wa:irasekrįxoʔš – You didn’t work – Weithioch chi ddim.

Is the fact that the Mandan word begins with ‘wa:’ and the Welsh with ‘wei’ a clue to their common heritage? ‘Wa:’ in Mandan is a negative prefix, whicle the ‘wei’ in Welsh is a mutation of ‘gwei’ which is the first part of the root ‘work’. No match.

Great care has to be taken in historical linguistics, not to see cognates where there are none. The only examples from Catlin that I see as similar are possibly: ‘head’ and ‘there is not’. Unfortunately for the latter, the morpheme analysis shows otherwise: in Mandan, ‘wįk’ is the verbal element, while in Welsh it is ‘oes’. The pronouns are extremely dodgy (as are grammatical affixes) because they tend to simplify themselves over time due to frequent use. It is unlikely that one will find a language with a common first person singular pronoun like ‘txwirpwyl’. As an example, the Welsh nominative 1st sg. pronoun is ‘i’, and the English is ‘I’. A match?  The pattern of evolution for these pronouns is: Welsh ‘i’ is from ‘mi’ > ‘fi’ > ‘i’, whereas the English ‘I’ is from ‘ic’ > ‘i’. A better match is Welsh ‘i’ with English ‘me’, but this would not be obvious by simply comparing the two words without a full philological study. This is moot as the Welsh and Mandan pronouns don’t match anyway, and contrary to Catlin, Mandan does not have separate pronominal prefixes for masculine and feminine.

If you are a firm believer of Madog’s voyage to North Dakota, please do not email me with additional proofs or complaints. I am a Welsh speaker, I have met a Mandan learner, we could not communicate even a little bit. It is my hope that the history of the Mandan people can be respected for what it is, I don’t imagine they appreciate their oral history being relegated to second class by a European myth. There have been enough prejudicial theories against Native history: from the foreign builders of Cahokia to Egyptian influence in Maya writing. At least (in my opinion) the Welsh-Mandan story can be dispelled by solid linguistic evidence.

This is not to say that Madog did not make it to North America in the twelfth century, but simply that the Mandan language is not in any way Welsh. Many historians believed that the Norse never made it to North America, and that their stories were just that, stories. Now that archæologists have dug up L’anse aux Meadows, we can be sure that the Vikings were not just making up amusing tales. Perhaps real evidence will surface, showing that Madog did in fact land in North America; that evidence has not and will not be found in the Mandan language. Still not convinced? My advice is learn to speak Welsh, and then try and read this.

Update

I have seen another unreferenced list of apparent cognates between Mandan and Welsh.

I don’t know what language it is, but it’s not Mandan. It has no letters d, f, th [θ], nor would there be any reason to have a Mandan vowel y. Furthermore, there’s no hint of the nasal vowels which are so common in Mandan. Another false lead.