I spent some time trying out different ways to achieve palatalization. Here's basically what I was thinking:
"It looks like the development of palatalization was pretty straightforward in Greek, but if I change the phonology to be closer to Swedish, then that changes the available sounds produced by palatalization. I think that motivates some limited changes in voicing for certain consonants. ... Okay, it's a little tricky to come up with a satisfying derivation for this that fits with everything else."
I've come up with two basic ways to try to resolve this. The first felt excessively fine-tuned, but I like the second. I can't guarantee that I won't come up with a third that I like more. Anyway, I'll stick with this second idea for now, and try to firm things up.
Here's what's missing from my plans: lots of things relating to stress, such as vowel length. Beyond that, I basically know how I want stuff to fit together from a phonological perspective. The biggest unknown is that I'm pretty sure I want some kind of vowel sound to be present in the protolang and get dropped, and I'm still deciding what kind of sound, and when this happens in relation to determining stress.
Anyway, I'm going to publish this, and then try to do some prep work for something else that probably won't have a blog post associated.