Real Spellers

Great work Heather.

You've done great scholarship here. You've shown your hypotheses, and identified what you do and do not yet understand! Now that you know you are looking for evidence of and <-ost> suffix, if it is out there, you will be likely to uncover it one day! Until then, you can sit with a working hypothesis of <gnost> as a base -- even though you have a well-founded hypothesis that it can be analyzed more deeply. With that hypothesis, I can make these word sums (all of which share the same Greek root):

dia + gnost + ic → diagnoatic
pro + gnost + ic + ate → prognostic

And also...
a + gnost + ic → agnostic

Since <dia->, <pro-> <a-> and <ic> are all affixes we can easily prove in other word families, we can safely sit with this analysis and treat these words as etymological relatives with words like <cognition> and <gnosis> with a <gn> base for which you have evidence that you understand for all the other affixes:

co + gn + ite/ + ion → cognition
gn + ose/ + is → gnosis

If/when you find evidence for <-ost>, that would be the time that you can treat <diagnostic>, <prognostic> and <agnostic> as members of the same orthographic morphological family with the base <gn>. Until then, there is no hurry. It is fine -- and totally valid to say that all of these words are in the same etymological family ("extended family"). If/when you find the last piece of evidence that you need, that's when you can safely conclude that they area actually not just in the same "extended family" (etymological family) they are actually in the same "immediate family' (same orthographic morphological family).

I hope that you are proud of the great learning you've already done (and shared). Finding this last piece of the puzzle is not actually something to be in a hurry for. Perhaps the richest lessons available from your hard work is recongnizing the difference between "incomplete analysis" (totally valid, but not complete analysis) and "false analysis" (invalid analysis that violates the morphological structure).

Stopping your analysis as <gnost> for the words you can't yet prove the <-ost> suffix is the responsible, scientific conclusion that is an invitation to later learning. If you go deeper to <gn + ost>, you run the risk of false analysis that will have to later be undone to move forward in understanding.

This also helps reinforce that the actual goal of any orthographic analysis in SWI is NEVER really about understanding the specific word that starts the investigation. The actual target is deeper understanding of how the orthography system works, and how to become a better orthographic scientist.

In those two goals -- you have succeeded with flying colours. But to have that success REQUIRES not going deeper than you can prove!