This episode discusses how “images” enter the mind; it does a good job at to deciphering what our dear poet meant, however it falls a bit short in taking into account something I now consider a fact (until sufficient proof convinces me I'm wrong, in which case I shall happily change my mind):
In some ways, ancient minds worked considerably different from ours.
On one hand, there's computer-assisted archeology to back this up; here's a pop-sci TED talk which does a very good job at explaining it. Please do watch it to understand this post. On the other hand, there's what we can extrapolate from people who grew up among animals (“feral children”): they will not only behave like the animals they grew up with (in terms of locomotion, facial and vocal expression, food, social behaviours, …), but will also (as far as can be told) think how the animals think to a large degree, and in either case will be unable to “become normal” after a certain age (unable to acquire speech, to use tools, …). Then there's what we can extrapolate from people who grew up in various shades of isolation (criminal neglect, human trafficking, …): During the respective critical period many behaviours can be instilled into them, making them extremely complacent, highly dissociative, or otherwise “useful” to their offenders. And finally, there is what we can extrapolate from solitary confinement (isolation of adults): Otherwise healthy humans will start to develop odd idiosyncratic behaviours, such as naming objects and speaking to them as if they were human, speaking out loud to imaginary friends or memories of friends (in psychology, this is called a fictive heterodialogue); if they have a TV, they might consider recurring TV actors / show hosts to be their friends; with increasing sensory isolation, pseudo-hallucinations (hallucinations which are recognised as being unreal) will begin to appear.
I can assert the truth in Martin Buber's statement that “The human self forms in encountering another self“, as I know even the extension to be true: Without another self, one's self dissolves. We need another self to define our boundaries, and we need interactions, more specifically, we need conscious reactions (mechanical/mindless reactions won't do) to define our content (what the self is filled with; our values and ultimately our identity; the who-I-am as opposed to the where-I-begin/end). In such a situation, drawing on fictive heterodialogues helps slow this dissolution immensely. The more progressed it is, the more malleable we become – no matter whether out of our own will, out of someone else's will, or out of our subconscious drive to adapt and survive. The older we are, the lesser these effects will be, but I'm quite certain they never fully vanish. Without the non-stop reality testing provided through social interaction (next time you speak to a friend, casually sprinkle in a most outlandish statement to see it in action), with increasing time/intensity, our thoughts can go off in all kinds of directions. We know this effect to be true not just for individuals, but for relatively isolated groups of people, too, as is illustrated in the Salem Witch Trials, the people of the Easter Island who (possibly) caused their own extinction by logging trees for cultish purposes, or the North Sentinelese, who collectively fear foreign humans very, very much; what they all have in common, is that the “hive mind” of those groups of people is isolated and cannot readily interact with outside “hive minds” to do a reality test, leaving the entire group prone to diverge further and further what is real – even to their own peril. We can find similar reasons (isolation, …) for other irrational convictions (such as cargo cults), in groupthink phenomena, and in modern-day algorithmic filter bubbles.[1]
Now, to circle back: All of this shows the immense plasticity of the human mind and the overwriting influence of culture on it, and while Ancient Greeks did not live in utter isolation, their groups (eg minor islands) were sometimes much more isolated compared to the ubiquitous global communication of today. Together with the statistical computer-linguistic proof (→ TED talk referenced above) it is therefore fair to think of Lucretius' explanation of “images” like that: Epicurus' model tried to explain the “images” without having as introspective a mind as we have (but he already had much more introspection than Homer); the “images” would therefore have felt more like pseudo-hallucinations to him, which explains why they were of such importance to begin with, and why they required such careful, detailed explanation: to clarify that they're not messages sent to us from Olympian Gods, or muses. At the same time, they couldn't have come from “his own mind", because they felt distinctly foreign (like nightmarish fever dreams do), possibly even ego-dystonic at times, and he didn't necessarily conjure them up consciously, rather they might have happened to him in a subconscious manner.
Quick thought 1: This might also be why the geometricians were so happy to adopt the idea of a plane of ideal forms – they might have experienced such a plane, when pseudo-hallucinating geometric shapes after having spent all week drawing figures in the sand…
Quick thought 2: The introspective mind being a new development then might be why, personally, I feel as though their writings are somehow clearer, fresher, less cluttered – without two millennia of introspective cultural baggage, were their minds less tainted from the many layers of paint that are glossed over our selves, having grown up in modern society?
[1] Footnote: The lack of reality testing in the relatively isolated filter bubbles are the explosives at the core of the bombs that are so-called “social networks”; dangerous, but not by itself. The fuse to light it up is to be found in the imperative to generate an income, hence grow the user-base and absorb as much user-attention as possible (to sell ads) – this is achieved by designing apps such that they bypass the evolutionarily new, thinking parts of our brains, and instead speak more directly to our evolutionarily older, more primitive, instinct-driven, emotion-based brains.[2] It is these "back brains" which do irrational things such as scroll through Facebook at 3am on a week day or get hooked on five hours of video footage from the deep oceans. The two characteristics are biologically related: by designing an app which gets people hooked, they have to design an app which bypasses the rational, thinking brain, and by bypassing that, the users stop reality-testing what they just saw. Without reality testing what they saw, they are easily drawn into the parallel world of a filter bubble, which is algorithmically isolated, and thus self-sustaining. Ultimately, this leads to a fragmentation of society along the lines of these filter bubbles, hence to the inability if the society's “hive mind” to form a consensus (the “hive mind” is fragmented), which in turn can more easily lead to civil unrest (the same effect – trying to reach a society-wide consensus without all social groups having a say in it – is what brought about the French revolution).
[2] Footnote to Footnote: We all know this "back brain" effect: When our neighbour rings our bell in the middle of the night yelling "Fire!", we'll get out immediately – our "back brain" acted instinctively to save our life. Only once we're out on the street does our "front brain" realise: "I cannot see any fire. There's no smoke. No smoke alarms went off, either. My neighbours eyes look suspiciously red and glassy. He's known to be a junkie. This was a false alarm, and, darn it, I forget my keys!" We also know it as an “auto pilot“ effect, which can be specially intense when we're too tired to engage in conscious thought or when we're in a particularly stressful event (such as immediately after an accident).