All we can say for sure is that the "shanty" pronunciation was in use by 1858. Nordhoff's is unclear.

What is unclear about Nordhoff? (*Playing Devil's advocate here.*)

It's technically true that the digraph /ch/ can be used to represent more than one sound. So, it's not 100% clear, but I guess "unclear" to me connotes more than "not 100%". I find no suggestion *in* Nordhoff that it should be read other than "church". There's good sense in thinking that if he presented it as /ch/ in such familiar English words as "chant" and "chanting" (which is especially English morphologically) that to suppose other than "church" is overthinking.

So why would *we* every suspect "shush"? I think it's only because we have learned to pronounced /chanty/ that way, through later writers and received pronunciation. On the other hand, /chant/ and /chanting/ are earlier (within the few decades prior to Nordhoff) in evidence in English-language reference to work-songs, both of sailors and non-sailor African American workers. I don't think we're talking about those things being pronounced as "shush". Why privilege what came after rather than what came before?

What makes Nordhoff notable in the historiography is that his is the "first" where /chanty/ (as opposed to /chant/, i.e. with a /y/) occurs, albeit embedded in /chanty-man/. Yet, I don't think that distinction is enough to treat it as a kind of ground-zero. Nordhoff is talking about the songs as /chants/, but we've already got earlier people talking about /chants/ with a similar meaning. I don't regard those others as unclear. Again, perhaps technically we can call them unclear, but I'm inclined to read them as typical English.

What to me is unclear is how Our Word was pronounced by everyone and in all places, and Nordhoff *contributes* to the overall lack of clarity, but that is something different than to say that Nordhoff was unclear.

Did not the earliest usage of the word occur in the Gulf Ports?

My interpretation is that "chantyman" (whatever spelling) was in use prior to "chanty" (whatever spelling) and that it was used by stevedores before sailors. At least, in use to a significantly greater degree. Some of those stevedores were in the Gulf ports. I theorize "chanty" being a "back derivation" from "chantyman". I'm happy to think of "chanty" as a simple diminutive form of "chant" as well, per Lighter. However, if I have to choose, I lean toward my theory for the reasons expressed in my article.

But I would not pinpoint /chant/ to either the Gulf ports or stevedores. Rather, those places seem important for /chantyman/ (from which, again I derive chanty—a later name for the thing earlier called chant). It's a subtle point, I guess.

French/Creole being a language of the Gulf ports, that offers some reasonable fodder for speculation (along with the complicated French and English mix in some Caribbean islands) about how the "shush" pronunciation got in there. In other words, its useful only insofar as we are grappling with "shush", for which I find an origin in Standard French to be less satisfying than either a Creole linguistic environment or some random sound shift within English.

(Scots "chantie" and French "Chantez!" via "shanty-man," or, more plausibly, English "chant" + "Chantez!")
I've never understood the /chantez!/ thing, nor the /chanter/ thing (though the latter may just be people wanting to state the verb infinitive as some kind of etymological formality?). I guess they are supposed to be accounting for the /y/ ending? Yet in French we already have the perfectly good word /chant/, which was used in French for work vocalization (and is used now as a gloss for chanties, "chants de marins"). And /chantie/ sounds to me like an equally plausible French diminution of /chant/. English /chant/ and French /chant/ are more or less equivalent.