Q: What about the signs where you have some very convincing thoughts, the trader and so forth. The logical basis that traders would have to be represented in some way on seals which were meant for trade makes sense. [See Mahadevan's "Indus Script Dictionary" for his speculations]
A: It is a moot question whether the Harappans had castes like we have in later India. But they certainly must have had occupational groups, priests, scribes, traders, warriors, why not? Perhaps they were not very rigid divisions. In any case, even if these terminal ideograms represent different groups, we must remember that the ideograms themselves are combined - for example the symbol of a bearer is combined with that of a jar, and that is followed by a harrow sometimes. So they could not have referred to exclusive caste groups like we understand the term. The possibility is that originally these were all mere phonetic symbols, but later the groups whose names used these symbols most, they only had surviving symbols with the loss of the language, so the symbols became mythical symbols or representations of their own cult objects. It is a jungle really, very difficult to trace out the environment in India which is not only multi-cultural, multi-racial and multi-linguistic.
The Dravidians and the Indo-Aryans interacted with each other and we have the Indo-Aryan languages with Dravidian features and we have Dravidian languages with Indo-Aryan features. Hinduism has both Aryan and Dravidian elements, not to speak of modern Hinduism also being influenced by Islam and its own native offshoots like Buddhism and Jainism. There are no clear cut parallels and what is more, the same symbol in the Indus valley could have taken multiple forms later.
The jar sign is one example: Kulda is both jar and the fire pit. So, predictably in later India, there are communities which claim to arise from the jar and others to arise from the fire pit. Now both have the same word and both could be two very different reflexes of an original common tradition. So, the claim can be made by students like me studying the survivals of Harappan culture in the later Indian traditions. They are no more than pointers rather uncertain, but nevertheless, if there is accumulation of evidence, if there are many pointers to a single fact, then one can go by it and try to find out where it leaves it. Some kind of a preliminary work to have a framework to have a hypothesis without which you cannot really start looking. If you are deciphering it, it’s upsetting to say that I decipher the script without any idea what the language is. In a logo syllabic script, it’s impossible. If you are going to use rebus then use a language because rebus is a pun, and pun are language specific. I know that Michael Ventris deciphered the linear “B” as Greek even suspecting it was so, but then he was dealing with a purely phonetic, syllabic language without any ideographic element. Now, that is not the situation with the Indus script. So, one has to have a hypothesis and then start. Discard the hypothesis if it doesn't fit with the facts.