Historical Linguistics Book

I often daydream about linguistics. Between making little discoveries of the vast borrowings between Persian, Arabic, Turkish and Urdu, and inventing my own language, I can spend the entire subway ride in the morning just thinking about how a truly solipsistic morphology would look in the verbs of a language which was only ever spoken in the first person. I have a passing understanding of linguistics, so this book was the perfect manual (liber manualis) for thinking about how the different scales and structures of a language interact and influence each other. The book works through various combinations of four levels of language : phonology, morphology, syntax and semantics, as I guess all linguistics does (news to me). It then uses colorful examples from World literature, and when I mean World literature I mean it in the way Bilge Karasu does "Literature is...the memory of language. I am not saying the memory of individuals, it's the memory of language.", to explain what the hell it's talking about.
It's the world and time wide jaunt to the Behistan inscriptions in Iran, to vikings writing on bones, to stuck up brahmins, to Akkadian slang. In one of the best passages, it talks about how both the family tree diagram and the venn diagram conceptualizations of language interrelationships are both wholly inadequate to show the complexity of their situations. A relic of when a language had object verb order gets lodged in a haughty adverb, norsemen get lazy with terminal vowels and lose the Indo-European heritage of a verb, the word Rhyme, from Rim, gets a weird spelling from the thinking of an erroneous folk etymology that it is related to rhythm. The vast continental sediment of human thought has the very language I am writing in produced from celtic place names, norman legal terms, Scandanavian pronouns, and Hindustani quirks.
The book is getting to me. I did not find many examples of the interaction of Arabic, Turkish, and Persian, but I can now call those things by their scientific name when I come across them. In my Arabic class I keep pointing out how the lack of voiced labio-velar approximant leads to much more recognizable forms when words were adopted into Turkish. Qahwah to Kahveh, baqlawa to baklava. Now I understand that the adoption of the Arabic verb for to guess "khamman" and to repair "3ammir" had their verbal nouns attached to make compound verbs in Persian and Turkish: tahmin etmek/ tahmin zadan and tamir etmek/ tamir kardan, and that this was a perfect example of morphological adaptation in borrowing. I want to stop time so that I can make a dictionary that compares semantic changes through borrowing where the meanings are extended from Arabic to Persian and Urdu. (since when does فیصلہ mean decision?!)
It also helped to reinforce the truth that even though I may think ways of speaking or ideologies about speaking are somehow unique in being contemporary, wincing when I hear "he be like crazy yo" or thinking that our language is devolving thanks to moronic pop culture, in fact this is the normal life cycle of our species of language on the colorful coral reef (how sad that climate change is taking that metaphor away from us) of language.
As I invent my own language, and think of better ways to express things, or more logical ways to use morphological derivation to add to its lexicon, I can see older traces of the "original" language underneath the form, even though it's only a few years old. It is a constant rewriting of a draft of a draft of a draft. It is a controlled experiment in historical linguistics. Although since it's never spoken it doesn't have to deal with phonological issues, or other people's input ;)