I’ve read that article before (might even have been because of a link Gabriel had here to that article previously). IMO it is a really nice article and a great compendium of sutta references for the four path stages.
It puts up a good argument, but as with many of these things it can really be hard to know for certain. IIRC the argument goes that references to sakadāgāmin and anāgāmin stages, particularly sakadāgāmin, are far sparser than for sotāpanna and arahat. However, it’s not surprising that there are intermediate stages found in the suttas. Naturally, if an almost (but not quite) arahant died, followers or mendicants would almost surely have asked the Buddhas about his/her fate. It would have been very surprising if no states between sotapanna and arahat were found in the suttas. There doesn’t seem originally to have been that much interest in the early sangha in the two intermediate states.
The articles argues that later there was more of an interest in these states and perhaps some of the body of doctrine was retrofitted onto these states (fetter models etc.). Plausible but hard to know for sure.
I’m somewhat less convinced about the paper’s conclusions on the sotāpanna. IIRC its argument mostly is that being sotāpanna was originally equivalent to just being a fairly devout Buddhist convert (having certain beliefs and following certain behaviour and a moral code) with notions like having a penetrating insight into dependent origination and the opening of the Dhamma eye being somewhat later developments (or maybe a different strand of viewers as mostly strongly represented in the SN). You could read that into certain passages but IMO there’s as many that would indicate otherwise. The reading seems to be that perhaps the SN’s sotapanna treatment was a somewhat later upping of standards re the stream enterer, but you could as easily take the view that maybe weaker statements outside the SN were a watering down.