"It must be something simple," he declared.
Roi paused courteously. "What must be, father?"
"Whatever underlies it all."
--[page 15]

The bulk of the book is devoted to these twin projects and they perfectly reflect each other; Roi has no scientific materials and is forced to work out everything from first principles while Rakesh can draw upon all of human knowledge and construct scientific measuring devices at will but, despite the radical differences between the two groups in terms of access to resources, both use the same mode of scientific thought; coming up with hypotheses, running tests and then working out the meaning of the data. This is an interesting concept for a novel.

A lot of SF, while claiming to be "about ideas" is really about sensawunda-inspiring technology such as vast habitats, star drives or some perhaps kind of gun that disassembles you at the atomic level and then reassembles you leaving you unchanged except for the unshakable feeling that your name is Frank. Incandescence does not try to instill a sense of wonder, nor does it contain any Big Ideas. Instead its focus is upon solving small-scale scientific problems. While this is clearly an audacious experiment on Egan's part it unfortunately makes for a rather dry read. Egan's prose is simple and unadorned and as a result his ideas are always clearly conveyed but, unlike tales of Big Dumb Objects and killer robots, reading about someone working out gravitational fluctuations or performing experiments using springs and tubes is actually quite dull. Clearly, Egan was sufficiently intrigued by Roi's world to put up some drawings and maths on his website but in truth, the amount of detail lavished on these chains of reasoning smacks of self-indulgence and is likely to cause all but the most obsessive of physics geeks to shift into "skim mode." However, what is most frustrating about the misfiring of these chapters is that the book's creative priorities are determined by them.

As I said in the introduction, it is something of a cliché to criticise Greg Egan for weak characterisation and plotting. Indeed, this is a cross that pretty much all Hard SF writers have had to bear. But to criticise Hard SF for these weak elements is also to completely miss the point.

A technically well-written book is one where the different elements that make up the book work together to support the book's "point." For example, most mainstream modern novels are deeply wedded to the social realist tradition of being about mundane matters such as George Elliott's Middlemarch being about education and class. Given such a subject matter, it is only natural that prose style and characterisation should be prioritised as they serve to aid the depiction of the situation whilst keeping it entertaining while the characterisation gives depth to the relationships that make up the text. Similarly, a mystery novel or a thriller is about a series of events and so it is important that the plot be well-paced and that it hang together in an interesting manner.

Works of Hard SF, by contrast, tend to be about speculation rather than people and that speculation tends to involve complex ideas that need to be communicated in as clear and efficient a manner as possible. To introduce formal innovation or complex and subtle prose into such a context would prove just as counter-productive as a network of limelight-hogging relationships and baroquely drawn characters. As a result, it is foolish to criticise Hard SF for its formal and human elements as writers of Hard SF intentionally push these aspects into the background.

Unfortunately, in terms of writing strategies this does amount to putting all of your eggs in one basket because should any of the foregrounded ideas fail to engage then you are left with a very dry book indeed. This is precisely the problem that Incandescence suffers from.

Egan's failure to properly marshal his resources and write to his book's genuine strengths is further exacerbated by the fact that many of the book's secondary ideas are genuinely fascinating and could easily have redeemed the book had Egan invested more heavily in them. For example, one of the book's main themes is the fate of the intellectual. Compelled by his nature to focus upon larger issues than those that pre-occupy most people, the intellectual can frequently be a strange and lonely figure. We can see this in Roi's initial encounter with Zak :

Ahead of her, a lone male stood idle in the middle of the tunnel. He wasn't begging for help, and as she drew closer Roi could see that he carried no seeds. A moment later she recognised something else in his appearance: the visibly labouring heart of someone who'd ventured well beyond the weight he was accustomed to bearing.
--[page 15]

"If you wanted to, you could wake us from our sleep "
--[page 261]

Egan's decision to focus upon one particular plot strand at the expense of all others is also evident in the fact that while the bulk of the book appears to be building towards a plot-line-uniting old school first contact scenario, it is never actually clear that the two series of events intersect at all. For all we know they could well have happened millions of years apart. Instead one plot line reveals some information that is important to the other plot line but the ramifications of this revelation are not explored in the least. They are simply allowed to hang there leading to the feeling that the book's ending does not work.

For all its hints of greatness and pleasing moments, Incandescence ultimately feels like a failed literary experiment. The counterposition of resource and data-poor scientists and a more data and resource-rich group is inspired as is the exploration of the social aspects of an intellectual lifestyle but unfortunately both of these promising ideas are starved of the characterisation and relationships that would have been necessary to properly sell these very human issues. Instead, the book is structured round a series of chapters which, while artistically adventurous (you rarely see sustained scientific argument in SF) are also frankly rather dull.