Settlement of the Americas

It is generally agreed among scientists that the Americas were first settled by migrations from Asia, across the Bering Strait during a period of greater glaciation and lower sea levels that permitted travel by land.

There are many uncertainties; conservative estimates have traditionally claimed that this settlement happened as recently as 12,000 years ago; more recently, earlier figures, even to 20,000 years ago, are more typically cited. The routes taken subsequent to the Bering crossing are not certain either. Some claim that migration occurred by boat rather than by land. There is a good overview of different positions at Wikipedia.

Geography of North America

This continent is of course divided by major geographical features such as mountain ranges, which tend to form barriers to cultural contacts, and rivers that may form either barriers or centers of activity. This can lead over time to the break up of language families and development of distinct cultural practices.

Geographical regions

Here is one breakdown of the physical features into specific regions. Notice that the High Plains may have served as a route for the southern movement of early populations.

Physical regions of the US

Focusing specifically on what is now the continental United States, some broader geographical regions can be labeled as shown. These physical contexts correspond to a large degree to culture areas, seen next.

Culture Areas

Somewhat along these lines, although affected by many other facts of geography and history, we find a number of broad areas in which cultural patterns – which might include food gathering and preparation, craft work, housing, marriage practices, religion, music, literature, etc. – are relatively similiar among different groups of people. This map labels various tribal groups within the areas.

These shared features are the result not only of similar living conditions (the cultural version of biological analogy), but also of contact and interaction over time (homology). Certain linguistic features can spread within these regions, which gives rise to a linguistic area that spans genetic relations; but it need not at all reflect the language families, since speakers might move to a new region and adopt much of the local culture without replacing their language.

Degrees of relatedness

We'll consider issues about language relatedness in more detail later in the semester, but it's good to have some overall idea of what a "language family" is before proceeding to discussion of diversity of sounds, word structure, etc. Above are some basic terms, listed according to increasing size and decreasing closeness of relationship.

Other terms for stock include phylum, superfamily, macrofamily. In work on American languages, stock or phylum is typical.

Some languages spoken by a (very) small population have no significant regional dialectal differences within the speech community.

US language groupings (Sturtevant 1967)

The colors on this map, based on a scholarly conference in 1964, correspond to groupings of related languages, or rather languages that the attendees at the conference agreed were probably related; some of these "stocks" are not widely accepted today. The languages of Alaska are in a separate map.

The gray lines on this map reflect the culture areas, which in many cases quite different from the relations among languages. In other cases, though, a culture area and linguistic grouping align to some degree; notice the eastern edge of the Plains region lines up with the Souian family. This is natural because edges of culture areas often reflect barriers to movement or communication, which affects the distribution of languages as people move, or remain in an area. But at the same time, these barriers are not insuperable, and people do cross them when there is sufficient motivation – such as population pressures, tribal conflict, drought, better food sources, etc. This leads to a language family that spans multiple culture areas, such as Algonquian or Uto-Aztecan.

Click here for a much larger version of the map, where the individual language names can be read.

US language groupings (Goddard 1996)

The analysis in this map was first published in 1996 under the editorship of Ives Goddard, and is based on more conservative views regarding whether languages are related in the same family. In particular, certain proposed larger groupings such as Hokan and Penutian (in the Pacific states) are not accepted here. There are still a number of large or relatively large families, however, which are listed at the right of the map. White areas are either sparsely populated or of uncertain linguistic affiliation (the latter in the eastern US).

The Campbell book in the readings takes a similar point of view; we'll discuss these issues in more detail later in the course.

See a larger version of the original map here.

North American language groupings (Goddard 1996)

Here you can see the languages extending further north. Much of this is the large Algonquian family, but is also the locus of most of Na-Dene and Eskimo-Aleut.

Click here for a larger version of the original map, which identifies the individual languages.

Eskimo-Aleut family

This family, as its name implies is divided into two basic divisions. There are several Aleut dialects (native name Unangax̂ for the people and Unangam Tunuu for the language), spoken today by no more than 100 people. Most of the linguistic diversity and larger population is in the Eskimo family.

Eskimo-Aleut languages

The Eskimo family has two basic branches: Yupik in the southern half of Alaska and across the Bering Strait in Siberia (three main languages and their dialects), and Inuit elsewhere.

Inuit is further broken down into Inupiaq (mainly Alaska: see terminology) and Inuktitut (Canada). Kalaallisut is the native name for the Eskimo language in Greenland; it is sometimes also called Greelandic. There are thousands of speakers in each region, though the smallest number (3000) in Alaska.

Na-Dene family

This family consists mainly of the Athabaskan family, plus several smaller individual languages in Alaska and British Columbia. Notice that there are Athabaskan languages along the Pacific Coast (several small languages such as Hupa) and in the Southwest (Navajo and Apache). These are the result of population movements rather recently, around 1400.

The most promising, though still controversial, proposal of a relationship between American languages and their origins in Asia involves Na-Dene and the Yeniseian family of central Siberia.

Alaska in detail

This map shows the Athabaskan languages of Alaska and adjacent Canada (these are all the non-coastal territories), along with the three languages often classed with them.

• Eyak, which is clearly related.
• Tlingit, which is more distantly related to the Eyak-Athabaskan group.
• Haida, which has been proposed as part of the same family, but is treated as an isolate by more conservative analyses (including the Goddard map).

Isolates

An isolate is a language that cannot be classified into a family with any other languages; that is, it does not show systematic resemblances that are best explained by a shared historical origin. The Basque language (Spain and France) is a famous isolate: it has no demonstrable relation to other languages of Europe, or elsewhere.

Scholars sometimes differ about whether a particular language is an isolate. In particular, a language might not be closely related in a "family" with others, but might belong to a larger "stock"; thus this depends on whether one believes in the stock. Various attempts have been made to group Basque with other languages, some quite loony; none have sufficient evidence to stand up to scientific scrutiny.

Isolates in North America

All the individually labeled languages in dark gray on this map are isolates (by a relatively conservative analysis); their names are listed at the right of the map. Haida is number 28 on this map; some other languages of North America that are often considered to be isolates include Zuni (New Mexico) and Timucua (Florida). Some of the families listed by color along the left are very small, and many consist of just two or three languages (such as Tsimshianic, Kalapuyan, Alsean, etc.).

A common feature of proposals for various language stocks is that they place isolates within larger genetic relationships, so they are not truly isolates anymore (although one might speak of an isolate within a stock, meaning it has no close relatives). For example, if Hokan is real, then languages such as Karuk, Washo, Esselen, and Salinan would not be isolates, although again sometimes one speaks of an "isolate" within a larger family, if it is not part of a subgroup. Some other languages within the proposed Hokan stock belong to clear families, such as Kashaya in Pomoan, so they are not isolates under either assumption.

Algic family

The Lenape language, spoken where Philadelphia now stands, is part of the large Algonquian language family, spoken over a wide region of the US and Canada. Two languages of the northern California coast, called Yurok and Wiyot, together with Algonquian form the Algic family. The two California languages are sometimes called Ritwan, based on an earlier claim that they constitute a separate branch of the family; but, remarkably given the geographical distribution, there is not good evidence that they are more closely related to each other than to the rest of the family.

Algonquian languages

This image, part of the larger map by Ives Goddard, labels the Algonquian languages in more detail; they are shown in a rose-tan color.

Algonquian in Algic

Note that Eastern Algonquian is the largest branch of the Algonquian family. Lenape, here called Delaware, belongs to this branch and is fairly closely related to (some) Native languages from the Canadian Maritimes and New England to Virginia.

Lenape

Delaware Unami Lenape Lenni Lenape

The Native language originally spoken where Philadelphia now stands goes by several names:

• Delaware : a cover term for several different groups with closely related languages or dialects, shown in this map
  • named for Sir Thomas West, Lord de la Warr, who was the first governor of the Virginia colony, via the naming of Delaware Bay and River which were first explored by the English during his term
• Unami : the subtype of Delaware from this region, specifically Southern Unami; native wəná·miw "downriver people"
  • apparently from na·m "downriver" and w-...-iw "person from the specified place"
  • it meant something different further north, since "downriver" is a matter of perspective
  • linguists use this general term for non-Munsee speakers
• Lenape : the native term, ləná·p·e, from lən- "real" and -a·p·e "people"
• Lenni Lenape : the same word with added lə̀ni also meaning "ordinary, real, original"
  • used to distinguish one's local group from others in this general classification, but said to be "rejected as redundant by twentieth century speakers" (Goddard 1975: 236)

One of the last native speakers of Lenape was Nora Thompson Dean, or Touching Leaves Woman, who helped create a number of teaching materials. She died in 1984, but her voice can be heard at an online dictionary of the language. It has been reported that her brother Edward Leonard Thompson, who died in 2002, was the last native speaker in the US, although several speakers are mentioned in the Moravian band in Canada. Current tribal members in Oklahoma and elsewhere are working to revive the language; scroll down on this page for details. The language has been taught at Swarthmore College by Shelley DePaul of the Lenape Nation of Pennsylvania, descendents of Lenape people who were not displaced or who soon returned; students have helped produce various language materials.

Many placenames of Lenape origin are listed on this Wikipedia page, but the origins described may not all be reliable.

The Delaware Removal

Most members of the Unami and Munsee tribal groups were pushed out of their homeland by the encroachment of European settlers. Movement occurred in various directions and in complex stages, with ultimate result that the majority of descendents of Native peoples from this region live on land very far from their ancestral home. The largest body is the Delaware Tribe of Indians (or Eastern Delaware), which has a contested relationship with the Cherokee tribe and landholdings in Oklahoma. A distinct body in Oklahoma is the Delaware Nation (or Western Delaware). In Wisconsin, the Stockbridge-Munsee Community includes individuals with Mohican and Munsee ancestry due to an early movement of northern Munsee Indians into Mohican territory, with later displacement to the Midwest. Some Delaware descendents live in Canada (with three distinct recognized tribes) and there are tribal groups recognized by the states of New Jersey and Delaware.

Read more about the Delaware Removal.

Example: Kashaya

Turning to a Native language, we will consider Kashaya, spoken on the Pacific coast in the northern part of California.

Kashaya in context

This close-up shows the region with modern labels on a satellite map. The largest city nearby is Santa Rosa on the right side of the map, where the tribal office is located and where many tribal members live.

Pomoan languages

This map shows Kashaya with the other members of the Pomoan family.

Kashaya classification

The five levels of classification for Kashaya are shown here.

Kashaya dialects

The dialects within Kashaya are not very well attested, partly because they were not apparently very different to begin with, and perhaps partly because they had to some degree been minimized by consolidation of populations before in-depth linguistic study was conducted in the 20th century. Robert Oswalt, in his 1961 dissertation on Kashaya, describes the difference as "less than that between the English of New York and Chicago."

One difference involves the form of a "switch reference" suffix used in certain subordinate clauses. It takes the form /pʰi/ or /čʰi/ in standard Kashaya (depending on the preceding consonant), but is /hi/ in the northern dialect.

For example, Essie Parrish, the main consultant for Oswalt, grew up using /hi/ in the northern part of Kashaya territory (inland from the modern town of Stewarts Point, close to the site of the rancheria, or small reservation); but she switched to the standard version later in life as she had more contact with speakers from other areas.

Another known difference is that the central coastal dialect at Fort Ross simplified a double /ll/ that results from adding the locative suffix /li/ to the object suffix /l/. For example, Herman James was an elderly speaker who learned the language from his grandmother, and she grew up at Fort Ross, so he did not use the doubled consonant; but his daughter Gladys grew up on the reservation and used the standard form.

The structure of this word is /mo/ "run", /mul/ "around", /ya/ visual evidential, /l/ object, /li/ locative. (Three ways to think about the Fort Ross form: the geminate /ll/ is reduced phonologically to /l/; the object suffix is omitted; or the locative takes the special form /i/ in this context.)

Pomoan cognates

You can see by comparing a few vocabulary items that the Pomoan languages resemble one another quite strongly. (Central and especially Southern are closely related to Kashaya, as you can probably see.) These cognates are related words in different languages, which are hypothesized to derive from a single ancestral word, as spoken in the language we can call Proto-Pomo.

These forms are taken mainly from McLendon (1973). The forms in Proto-Pomo are generally fairly similar to Southern Pomo and Kashaya, which preserve a lot of the old distinctions that have been lost in other languages of the family.

A hyphen means the word is a bound form and may need to occur with another element to complete the basic meaning; for example, Southern Pomo "stream" is biʔda-kʰa, where kʰa is "water"; and the Eastern Pomo element can be identified in bi·dá-mi "stream" and bi·dá-w "downstream".

The Pomoan family tree

The members of a language family often can be grouped into subfamilies, or branches (of the "family tree"), based on the degree of similarity. Here is Oswalt's subgrouping of the Pomoan family.

This tree indicates that the people who eventually spoke Eastern, Southeastern, and Northeastern Pomo must have split off from the main population relatively early (there is reason to think Southeastern split off first), whereas the remaining people, who later spoke what we call Kashaya, Northern Pomo, etc., must have remained a cohesive group for some time, and then later split up in stages. During each period of separation, the languages naturally changed over time and therefore become more distinct. If a group remains cohesive for a period, their forms of speech will share these changes, and this is what it means to be more closely related.

The Hokan stock

A "stock" (or "phylum") is a collection of language families claimed to be a single super-family of related languages. The evidence for these is typically rather weak, a topic to which we will return later on. But the classsification in the map shown above shows stocks such as Penutian and Hokan, mainly in the Pacific states.

In this detail, the green areas are the languages classified as Hokan; this includes Pomo(an), clearly labeled, as well as the Yuman languages, which are the Hokan grouping in the south – in Arizona and Southern California (and into Mexico).

Remember that the heavy gray lines delimit cultural areas and are not linguistic boundaries.

California languages

This map gives a clearer view of the various languages and (proposed) families. Most of these groupings (such as Athabaskan and Chumashan) are clearly correct, but remember that Hokan and Penutian are just hypotheses. One exception: within Penutian, the Miwokan and Ohlone (a.k.a Costanoan) languages are well motivated as a family called Utian.

Pomoan and Yuman

Although, like other stocks, Hokan is just a hypothesis, Langdon (1979) argued for a connection between Pomoan and the Yuman languages, based on various morphological similarities as well as possible cognate roots. Some examples are shown here, simplifying a few details. Some of the Yuman reconstructions have morpheme breaks, indicated with a hyphen.

You can see consonant correspondences such as b:p, š:ṣ, m:m. The vowels correspond fairly well also.

Campbell (2000), who takes relatively conservative positions on language grouping, finds the evidence promising but not yet persuasive.

The Amerind super-stock

Joseph Greenberg (now deceased) proposed that all the languages of the Americas other than Eskimo-Aleut and Na-Dene (both well-accepted families, barring a couple of languages they grouped into Na-Dene) belong to a single remaining superstock called Amerind.

The Amerind super-stock

Within this group are further divisions that are still large enough to be called stocks, including many proposed by other scholars previously (such as Penutian and Hokan again), but have been adopted and modified by Greenberg and his collaborator Merritt Ruhlen.

The map above shows their classification of Amerind languages in North America; this proposed superstock also includes all the languages of South America. The "families" here are already large stocks, and none are well accepted, much less the grouping of all of them together. Hokan has been enlarged here, as a branch of Amerind, to include many languages of Central and even South America (Yurumangí, on the west coast of Colombia, which is barely attested).

"Penutian" within Amerind

As mentioned for Hokan, the groupings within Amerind are often larger than these terms otherwise describe. For example, this map of Penutian includes many languages that most modern proponents of a Penutian family do not consider plausible, although these broader connections were generally proposed well before Greenberg adopted them. Thus Zuni (an isolate), Gulf (mainly Muskogean), and Mexican Penutian (including Mayan) are not seriously entertained even by those who are currently trying to establish relations between, say, Tsimshian and the Oregon and California languages.

We will return to this topic later in the course.