This is more an artifact of how we typically define writing than anything meaningful about the act of communication itself. Graphical symbolic communication is tens of thousands of years older than what the article discusses. Writing as typically defined is a complete system for encoding verbal language using specific, formalized symbols. That's much more sophisticated and largely unnecessary for "most" human activities prior to the invention of large, hierarchical societies.
Proto-writing is still quite far along on the spectrum I'm talking about of contextually defined symbols. Lascaux and chauvet have plenty of examples generally agreed to be partially symbolic, just off the top of my head.
Can't say I agree, and I suspect you probably don't agree with the implications of that definition either. As a hypothetical example, do you think any of the following aren't writing:
* a script which can represent taxes, histories, and religious texts but not the full range of verbal expression
* programming languages
* emojis
The first of these is an actual scholarly debate about whether Aztec script can be considered "full" writing or merely proto-writing.
