When I get two articles (or more) on the same day about a non-election issue, I take notice. Today, I found two articles in my newsfeed on how quickly immigrants to the United States learned English in the past. Greg Laden's post goes into a discussion of his own encounters with this in Boston in the early 1980s, and one can also easily think of how a similar encounter might have taken place in the Italian market setting of Rocky's Philadelphia in the mid-late 1970s (or even today in New York's, San Francisco's and Los Angeles' Chinatowns).
The PhysOrg brief of Salmons' work parallels Greg's later points (since Greg's post seems to have been motivated by Salmon's article) about how German was actually the dominant language in parts of Wisconsin during the early 1900s. One key point taken from the Salmons article is that the 1910 census of Germantown, WI showed that 43% of US-born residents only spoke German, and this sizeable percentage was also seen in other towns and counties. Furthermore, German newspapers pervaded the region until the 1940s (at which time the newspapers were usually consolidated into larger ones), 100 years after Germans settled the region.
I agree with Greg's point that the myth of previous generations trying to learn English once they got here is just that: a myth. The fact that US-born primarily-Chinese speakers in cities' Chinatowns still occupy a sizeable population is an example of this trend continuing. Just because English happens to be a language spoken by a national majority does not mean that it should be an offical language. As a friend of mine pointed out, you cannot have freedom of speech while also enforcing an offical language.