The story of Mayan
01
From Friars' Grammars to a Reconstructed Ancestor
Spanish missionaries were the first outsiders to compare Mayan languages. From the 16th century friars wrote grammars and dictionaries of Yucatec, K'iche', Kaqchikel and other languages so that they could preach in them, and they noticed that the languages of the Guatemalan highlands resembled one another and resembled Yucatec. By the end of the 18th century European scholars cataloguing the languages of the Americas had also noticed that Huastec, far to the north on the Gulf coast, belonged with them.
In the 19th century the comparisons became more systematic. The Swiss physician and ethnographer Otto Stoll, in a survey of Guatemala's languages published in 1884, sorted the highland languages into groups that are still recognisable in modern classifications. No one since has seriously doubted that the Mayan languages form a family. What remained was to prove it in the strict sense and to work out its internal structure. Norman McQuown's classification of 1956 and, above all, the work of Terrence Kaufman from the 1960s onwards did this. Kaufman established the regular sound correspondences that link every branch, reconstructed the sound system of Proto-Mayan, and proposed the four-branch classification most work still follows.
The evidence is of three kinds. There are regular correspondences in basic vocabulary, of the kind shown in the table below. There are shared numerals and body-part terms. And there is shared grammar: every branch marks persons on the verb and the possessed noun with two sets of prefixes and suffixes that are recognisably cognate from Huastec to K'iche'. Since the 1980s much of the comparative work has been done by Maya linguists themselves, trained in Guatemalan programmes such as Oxlajuuj Keej Maya' Ajtz'iib' (OKMA), whose data feed into Kaufman's etymological dictionary of 2003.
02
A Highland Homeland?
Kaufman placed the Proto-Mayan homeland in the Cuchumatanes mountains of north-western Guatemala. The main argument is linguistic geography: the Q'anjob'alan, Mamean and K'iche'an branches meet in and around these mountains, so the family's greatest diversity lies there, and the most economical history has the languages spreading outwards from that point. Kaufman also argued that the reconstructed plant and animal vocabulary suits a mountain setting, though many of the species concerned also grow at lower elevations.
The reconstructed vocabulary is clearer about the economy than about the landscape. Proto-Mayan had words for maize, the grinding stone, chilli and a range of other cultivated plants, so its speakers were farmers. Kaufman dated the break-up to around 2200 BCE by glottochronology, which counts vocabulary replacement and assumes a steady rate of loss; most historical linguists treat such dates as rough guides at best.
On Kaufman's account, Huastecan left first and ended up on the Gulf coast of Mexico, while the ancestors of Yucatecan and the Cholan languages moved down into the lowlands during the 1st millennium BCE. That fits in broad terms with archaeology: settled village life in the Maya lowlands is attested from around 1000 BCE, for instance at Ceibal in Guatemala, and grew into the Classic civilisation.
Each step is open to question. Greatest diversity does not always mark the homeland, since later migrations can crowd languages together or erase them. Some archaeologists find it hard to see a highland population behind the earliest lowland villages, and the date of Huastec's separation, and the route it took, are uncertain. Ancient DNA from the Maya region is still sparse and has not yet been tied to particular languages, so genetics cannot settle the question.
03
Ergative Verbs and Glottalised Consonants
A Mayan language is recognisable first by its sound. All of them have a series of glottalised consonants beside the plain ones, written with an apostrophe: k and k', t and t', ch and ch', tz and tz'. The glottalised b' is usually pronounced as an implosive, with air drawn inwards. The Eastern Mayan and Q'anjob'alan languages keep the uvular stops q and q', made further back than k. Many languages distinguish short and long vowels, and a few have developed tone: in Yucatec, long vowels carry a high or a low tone that can distinguish words.
Word roots are typically a single syllable of the shape consonant–vowel–consonant, and the basic order of the clause puts the verb first, followed by object and subject or by subject and object depending on the language. In everyday speech, especially in the larger highland languages, subject-first order is common.
The grammar marks persons on the verb rather than on independent pronouns. One set of prefixes (Set A) marks the subject of a transitive verb and also the possessor of a noun. Another (Set B) marks the object of a transitive verb and the single participant of an intransitive one. This pattern is called ergative. In Kaqchikel k-in-a-ch'ay-o 'you hit me', in- is the Set B marker for 'me' and a- the Set A marker for 'you'. The same Set A prefixes mark possession: ru-kej ri achin, 'the man's horse', is literally 'his-horse the man'.
Other widespread traits include a distinct word class of positional roots describing how things lie, sit or hang, numeral classifiers in many languages, and counting in twenties. The Classic Maya wrote in a logo-syllabic script; since the 16th century the languages have been written in Latin letters, most recently in the unified alphabets adopted in Guatemala and Mexico.
04
What Is Still Argued Over
The existence of the family is not in question, but much of its internal structure is. Almost everyone agrees that Huastecan was the first branch to separate, and that K'iche'an, Mamean, Q'anjob'alan and Cholan-Tzeltalan are each real groups. The intermediate nodes are less secure. Western Mayan, which joins Cholan-Tzeltalan with Q'anjob'alan, rests on a small number of shared changes, and Q'anjob'alan also shares features with its Mamean neighbours. Several of the innovations that link the lowland languages, Yucatecan and Cholan-Tzeltalan among them, may have spread between neighbouring languages rather than descending from a single ancestor. Danny Law and others have argued that long contact between Mayan languages makes it unusually hard to separate inheritance from borrowing, so that the family tree may be partly a network.
The language of the Classic inscriptions is a second debate. Stephen Houston, John Robertson and David Stuart identified it with an ancestor of Ch'olti' and Ch'orti'. Others, including Alfonso Lacadena and Søren Wichmann, have pointed to regional variation in the texts, with Yucatecan features in the north, and prefer to see a Cholan prestige language written with local colouring.
Dating is a third. Kaufman's figure of around 2200 BCE for Proto-Mayan comes from glottochronology, and other methods give dates several centuries apart; none has independent archaeological confirmation.
Finally, proposals linking Mayan to other families remain unproven. Norman McQuown's Macro-Mayan hypothesis of 1942 joined Mayan with Totonacan and Mixe-Zoquean, and Ronald Olson proposed in the 1960s a link with Chipaya of Bolivia. Lyle Campbell and Kaufman concluded that the evidence for such connections was insufficient, and that some of the resemblances with Mixe-Zoquean are loanwords. Most specialists today treat Mayan as a family with no demonstrated relatives.