We should thank James Castellan for calling our attention to the Philip Sutton article (https://www.nypl.org/blog/2013/07/02/name-changes-ellis-island) since it plays a major role in propagating the meme that involuntary name-changes during immigration (there are similar narratives for other ports of entry) weren't possible. Some of the items on the list provided us by Barbara Mannlein a few weeks ago cite this article; others, not on her list, do as well. This is going to take some time, so bear with me.
There are several basic issues. Immigration procedures were even messier than Bob Bloomberg suggests. Congressional debate on HR 15442, which became the Naturalization Act of 1906, addresses actual fraud in the process involving agents of the Immigration Service. IS employees benefited from selling "duplicates" of immigration documents to an organization that then resold them, primarily in Italy. Other IS agents manipulated the naturalization process to enable immigrants who might vote for their party to become citizens faster. These were criminal acts, even then, so people who believe such agents would be meticulous about not changing immigrant names, which wasn't a violation of law at all, bear the burden of proving their beliefs. Anyone who wishes to check this out is free to start reading the Congressional Record.
The Sutton article raises key questions about the validity of the meme itself. He provides three "proofs," none of which can stand up to scrutiny:
Sutton's first "proof" begins with "Numerous blogs, essays, and books have proven" that involuntary name-changes could not have occurred. Such claims are standard in "proofs" of the meme, but while they invariably fail to specify where such a proof can be found, Sutton quotes at length from Vincent Cannato's American Passage: The History of Ellis Island including the statement "Nearly all ... name change stories are false." "Nearly all" implies that some such stories are true, which would disprove the "no involuntary name-changes" meme.
I read Cannato's book, which includes a statement of the meme and endnotes that identify his sources, but neither the book nor the references cited contains a proof. I then contacted him to learn whether "Nearly" meant what I had supposed. He responded that the conclusion that involuntary name-changes could not have occurred is based on what actually happened in the Great Hall at Ellis Island. He hadn't studied this issue in detail (his research was on the role operating Ellis Island played in establishing the American approach to bureaucracy); his "Nearly" was simply a hedge against the possibility that one or more cases might emerge where such a change did occur. Citing the meme, or citing a source that cites the meme, is hardly a proof.
Sutton's second "proof," involving the "One That Was," is a reference to the case of a Mary Johnson who chose to re-enter the United States dressed as a man and using the masculine name Frank Woodhull. Mary had adopted this name and identity after her initial entry from Canada, some thirty years earlier, when there were no federal immigration controls, because the job opportunities available to men were much greater than those open to women. Frank/Mary was referred for a medical inspection while re-entering, and recognizing that competent medical personnel would probably realize she was a woman, gave herself up. She went before a Board of Special Inquiry, which allowed her to enter. Immigration officials changed "Frank Woodhull" to "Mary Johnson" on the arrival document.
Contact with the Historian's Office at USCIS revealed that they believe Sutton has misunderstood or misrepresented what happened here: It was the listing for Mary that was changed rather than her name; she continued to live as Frank Woodhull. By virtue of having lived in the United States for thirty years, she appears to have known she was free to call herself whatever she wished. One of the arguments made to me when I first heard the meme was that Americans were free to do so. Making this a "proof" of the meme requires establishing that immigrants who had never been to America shared Mary/Frank's understanding of American law. Not likely.
Sutton's third "proof" is a claim that there was no contemporary discussion of name-changes, which is a form of the logical error known as an argument from ignorance (The best known of these is "Absence of evidence is evidence of absence," known to be false for centuries.) but then goes on to discuss an article he found in an entertainment column in The Washington Post from April 10, 1944, that does so. That article reports that a musician named Harry Friedman was reverting to his pre-immigration surname, Zarief. Sutton appears to have totally missed the context for this article: Mrs. Friedman/Zarief had just given birth to quadruplets and publicity was good for his career.
Sutton claims there are no other such items, but I found one, by sheer luck, dating to 1897. What would a comprehensive search of all media, including those in languages other than English, reveal? Sutton doesn't know, so this claim doesn't prove anything either.
The upshot is that the gold standard "proof" of the meme doesn't actually prove anything. It is built of claims that represent either misunderstanding or misrepresenting the evidence. The dozen or so other "proofs" of the meme I have found share this weakness, and some are credible only to true believers: One such adherent told me that whatever the law said, immigration officials would ignore it to do what genealogists would want a century later...
There is a mechanism that would lead immigrants to believe their names had been changed at Ellis Island, or other ports of entry. This mechanism is included in an earlier abridged version of my study of this topic as "Involuntary Name Changes: The Real Story," in Avotaynu, Vol. 34, #1, Spring, 2018, p.~34.
Contrary to Joel Weintraub, I believe the real issue here is whether the "No involuntary name-changes at Ellis Island" meme has led to the abandonment of narratives that contained genealogical information not available through any other means, and thus to its loss: You can't tell people that their narratives are fake without leading at least some of them to abandon them and thus prevent their transmission to a subsequent generation. Since stopping their transmission was the goal of advocates of the meme, they cannot now claim that there has been no such loss, unless they can prove that all the narratives are indeed false. But they haven't even tried to do that; instead, we have the meme. To make their point, they would also have to prove that narratives long-since abandoned were also false, but there's no obvious way to even identify them.
I note that there has been no systematic effort to disparage any other type of family narrative. Each narrative, including the name-change ones, should be checked out and not abandoned unless or until it has been proven individually to be false.
This meme has been effectively and aggressively marketed, which is fundamentally different from being proved, but that puts it on an equal footing with breakfast cereals. To the best of my knowledge, "involuntary name-changes on immigration" is the only area of intellectual discourse in which one side demands that the price of not being ridiculed is abandonment of the bulk of the evidence that proves that that side's beliefs are wrong.
What remains is to figure out why this meme emerged. I have requested a document from NARA that may contain, or point to, the answer, and as I suggested early in this thread, it would be wise to refrain from making additional unsupportable claims until it can be analyzed.