There are several sources for the Hanis language (and somewhat fewer for the Milluk and Siuslaw/Umpqua languages, although there is a lot of overlap – some linguists who worked on Hanis also worked on one or both of the others). The records with the clearest phonology – meaning, those that heard and understood the sounds of the words as spoken by a native speaker – come from people who did their field work in or after the 1930s.
Melville Jacobs, an anthropologist who spent is career at University of Washington, worked with James Buchanan and Frank Drew in 1932. Buchanan seemed to know only Hanis (given that he spent much of his life among Siuslaw and Lower Umpqua, I would be surprised if he did not know something of the language, but he left no records behind indicating that he did so). Frank Drew knew Hanis, but he also knew a fair amount of Siuslaw – but he shared only a few Siuslaw words with Jacobs, and a scant handful of Milluk ones. He made some wax cylinder recordings of these two men also (some of these songs can be heard on http://www.hanis.org). Then in 1933 and 1934 he worked with Annie Miner Peterson to get texts and vocabulary in Milluk and Hanis. He also made recordings of Annie singing and telling texts in both native languages. Jacobs’ written records are clear and pretty helpful, so it has been one of the important sources for the languages.
J. P. Harrington, eccentric and gifted linguist, worked for much of his career for the Bureau of American Ethnology. In 1942, he worked up and down the Oregon coast interviewing several Indians, writing down with great phonetic detail words in various SW Oregon dialects of Athabaskan, Takelma, Tillamook, Alsea, Hanis Coos, Siuslaw/Umpqua, and a few words of Milluk.
In 1953, Morris Swadesh and his assistant Robert Ultan were working on a Penutian language survey. They recorded Martha Harney Johnson speaking several Hanis words, Laurie Hodgkiss Metcalf and her sister Daisy Wasson Codding in Milluk, and May Barrett Elliot, Howard Barrett, Clay Barrett and Billy Dick speaking Siuslaw and Lower Umpqua.
These are the best records we have for reconstructing the sounds of the words for these languages.
There are earlier records, but as we go back in years the less phonetically accurate they are, and the more we need to reconstruct the words. Luckily, there is a lot of overlap between the early records and the ones from the 1930s or later. But, there are a few words that appear in early records that never got re-elicited by later linguists like Jacobs and Harrington. So, reconstruction is all we have to go by.
In 1906, Henry Hull St Clair interviewed Jim Buchanan for Hanis words and stories, and got some vocabulary in Milluk from George Barney. A few years later, Leo Frachtenberg worked with Buchanan and recorded more words and stories, incorporated St Clairs work with his and published “Coos Texts”. Frachtenberg also wrote a grammar of Hanis. He also worked on Lower Umpqua and Alsea. St Clair and Frachtenberg’s records are not quite as phonetically accurate as Jacobs’ and Harrington’s work, but they are quite good.
It is going back into 19th century records that things get tricky.
For Hanis, there is a vocabulary of about 420 words that a man named George P. Bissell recorded sometime in the late 1870s or 1881. Using the Powell schedule of vocabulary prompts that the US Department of Interior printed to help the study of Indian languages, he recorded this list, plus a list of over 500 words of Lower Umpqua, in two undated manuscripts. We know very little about Bissell, and his training. Trying to compare his lists with later ones, he was struggling with the very different sounds of Hanis and Lower Umpqua, and it is difficult to use his vocabulary. Also, he wrote that he thought Hanis and Lower Umpqua were dialects of each other, which is not true at all. The languages are distantly related, but even at a casual analysis could never be identified as dialects of a common language, or even closely related.
But, what piques my curiosity of Bissell’s record is, he did record several words that were not recorded by later linguists. Can these words of dubious phonetic quality be reconstructed to include in a contemporary Hanis dictionary? Perhaps. Most of his vocabulary words turn up on lists of later linguists, and so their forms can be compared to Bissell’s. From there, we might be able to interpret Bissell’s symbols to figure out the words that never appeared again in later sources.
Bissell recorded gall (gall bladder) three times – gis´kún, gis kûm, and gĭs´kŭm. The reconstructed word is probably giskom – a short I like ‘bit’, and a schwa in the 2nd syllable (which we’ve been representing with the symbol ‘o’; see the pronunciation guide in the ‘About’ section).
Gizzard is klāint. And here we come to a challenge. Kl- could represent several sounds. For comparison, he recorded kläs as black, which is q’le’es. Nail he wrote as klâ put, which is really tlapit. Ground as kl lâ is really tl’ta or tl’da. So kl- could be ql-, q’l, tl, tl’, or even xl-. So gizzard could have any of these combinations. Even kl- though it should be noted that is a rare combination of sounds in Hanis, so that seems unlikely.
Silver he recorded as koˇs toˇl´la. The 2nd word is tala, money, borrowed from the English dollar. The koˇs part is probably xqas, white. So silver is xqas tala, white money. I drew this conclusion in part because of the resemblance (albeit not great) between koˇs and xqas, but also that in later vocabularies gold was referred to as hlkwilt, red. So if gold is the ‘red money’ it is not much of a stretch to see silver as the ‘white money’.
So, some of the words can be reconstructed with a fair bit of confidence. And some, like gizzard, are going to be tricky. If not, perhaps, impossible.