How do we know what side is the "top" of the cipher?

Good question. There are indications that the sequential homophonic substitution was applied to the plaintext from left-to-right, top-to-bottom. Though that is independent of the plaintext direction itself.

– The first occurrences of every symbol in a top versus bottom scenaro show a top bias. Similarly, the shortest string that contains all symbols is found at the top of the cipher (by User3000). More on this: viewtopic.php?f=81&t=4461&p=73275#p73275
– The 408 has "increasingly random cycles", if the 408 is turned upside down then it would become "decreasingly random cycles". Likewise the 340 has "increasingly random cycles". More at: viewtopic.php?f=81&t=4099
– My "Merge sequential homophones" solver in AZdecrypt scores the 340 at 22054.41 normal and 21782.64 upside down. And the 408 at 25594.80 normal and 23593.99 upside down. No n-grams used, higher score is better. viewtopic.php?f=81&t=3198
– From doranchak’s wiki: Cipher symbols seem to get larger as the cipher goes on. Row 20 seems larger than Row 1. (From Doc. Doc speculates the author wrote the cipher from top to bottom, and was tiring.) http://zodiackillerciphers.com/wiki/ind … servations

Has anyone tried turning it upside down?

We tried that and millions of other directional interpretations at: viewtopic.php?f=81&t=3196

We tried that and millions of other directional interpretations

LOL

Yeah, I figured! You guys that work on the ciphers are doing God’s work, my friend. The stuff you guys do makes me feel dumb just for asking questions. Keep up the good work!