A new Indus script decipherment has dropped. The language is Dravidian/Tamil. Link in following tweet.
Method: signs are assigned values of what the author thinks it looks like. The circle/oval is assigned "pa" acrophonically from pakal (sun). But not really, because the wheel sign is assigned "pallai" elephant. A fish sign are "cart" and another fish sign is "raft". I would say essentially that the assignment is random and some flimsy justification is added later. [line 75]
Features:
Composites: Several signs can be combined to make new words that are sum of their constituent words. In general this is a good sign but only if it can be applied universally. [line 200]
Non-verbatim: Many sign values need to be changed to avoid nonsensical readings. [L 185]
Syllabic signs: Acrophonic syllabic values [L 195]. This is also good but only if it works consistently.
Numeric signs: Not consistent. Some are numbers, some are not. This is bad [L 220]
There are a few sample readings. Some of them are plausible and many like "enormously mature and soft farmer" [L250] are meaningless.
There are 7 sample readings. Only one is long and reads
"very clever trapper velappan of the triple mountain sends by boat along the big river with care to the tiller of the land" [L395] This seal (shown below) must have been used to make the same impression several times.
Correctness: In a logographic system, the number of values are unbounded and therefore the decipherment is unfalsifiable.
However, we see absurd combinations based on the assigned values:
For example, the three-fertile-farmer
Readers can dig around indusscript. net and find more interesting combinations.
I will say overall, this is the best Dravidian decipherment attempt so far because so many signs are assigned and long inscriptions are read (even if it requires a stretch of the imagination)