Book review: ‘Mumur’ by Will Eaves

Murmur image

“I cannot help wondering if the real nature of mind is that it is unencompassable by mind, and whether that Godelian element of wonder – at something we know we have, but cannot enclose – may be the chief criterion of consciousness.” So opines the narrator early on in the latest terrific book from Will Eaves. Startlingly ambitious in its scope and form, Murmur invites us into a world of philosophical mathematics and artificial intelligence. What’s not to love?

Now when it comes to these topics, Eaves has touched upon these areas before – for instance, within The Inevitable Gift Shop. Yet here in Murmur he explores it with an astute intimacy from the perspective of an avatar, Alex Pryor, a character based on the father of artificial intelligence, Alan Turing.

It is not whether or not machines can think that is the main focus here; but rather, a potential inverse of the proposition – whether or not humans think like machines. Murmur is more concerned with the nature of human consciousness, how we come to be – whether we are pre-formed, destined to live pre-determined lives following a set of codes within our basic DNA, or if we are our own programmers (to stick with the computer theme).

As Turing himself argued in his seminal paper, Computing Machinery and Intelligence, when asking the question ‘can machines think?’, it is firstly of critical importance to “begin with definitions of the terms ‘machine’ and ‘think’”. Determining whether or not something possesses artificial intelligence is not based on empirical fact, but rather, decision – the decision of the human beings setting the frames of reference for any AI test (the computer can play chess; can fool a human into believing they are conversing with another human; etc.). That a machine may ‘pass’ such parameters does not necessarily mean they have acquired genuine intelligence. As Noam Chomsky has argued, conversing with a computer shows only that a piece of software can be programmed to breakdown the codes of our language and repurpose them (as it has been told to do so by a human programmer). This is not intelligence; but parroting.

Yet the notion of conversing with a machine opens up linguistic questions and challenges. Numerous pieces of research have shown that language not only shapes our culture – but also shapes and manipulates our personalities. Language programmes us, in that sense. With this in mind – and considering the subject of Eaves’s book – the Turing test, which has for so many years been the gold standard of measuring a machine’s intelligence, becomes even more central to the core of Murmur. By choosing to frequently adopt a conversational style within his writing, the reader must begin to question the formal structure of the novel, and their relationship with both the words on the page, and the characters within it. Are we, as readers, engaged in a Turing test of our own? Asked without directly being asked to assess whether we are in conversation with machine or man; or, more simply, whether we are able to assess for ourselves what does and does not have consciousness? Do characters feel, if their actions and thoughts on a page make us as readers feel? Are books themselves alive, if they contain within them what looks, feels and appears for all intents and purposes to be consciousness?

These questions of course invite further questions. For instance, is it mere coincidence that formally, there are times Murmur’s structure resembles some of the (at first) seemingly disconnected pieces of text – memories, questions, letters, and so on – that might be produced by some of the ‘AI’ writing programmes that have been developed in recent years? Coincidence perhaps; yet the fragmentary nature of the novel certainly asks us to think about the ways our own ‘intelligence’ – or consciousness – is structured.

We like to think of ourselves as straight thinking, coherent and logical beings despite all evidence to the contrary. There is no clearer feature of the mind than its willingness to construct wholes out of fragmentary parts. Our memories inevitably have gaps within them. Our focus can so easily be lost to distraction. Thoughts and memories pop up seemingly at random. A innocuous smell or sense of touch can make us involuntarily recall feelings and thoughts both good and bad; as well as those we have suppressed.

Life and consciousness are not logical (though they can of course be assessed and reviewed with logic). And this is one of the many things that Murmur does so well – it is, by its very nature, both an accurate representation of consciousness and human experience, as well as a thorough, logical analysis of these things. Through Alex Pryor, Eaves has developed a protagonist through which we may see these inherently complex ideas more simply.

This would be a triumph in itself; yet Eaves goes further – creating characters that are not simply tools through which we may explore high-level concepts, but through whom we empathise with, laugh with, and love with.

Perhaps this last part is the most important (as it so often is with a good novel). For all that the writing is excellent (as we have come to expect with Will Eaves); and for all that the book grapples with a veritable menagerie of ‘worthy’ ideas (there are so many more we could have discussed at length in this review); and for all that it provides another worthy voice to consider in the ongoing conversations surrounding artificial intelligence – none of these are really what the book is ‘all about’, or what readers should take away as being the most important aspect of Murmur. Because ultimately, what it all comes down to is that this is a novel about love. And it is the way in which Eaves presents this most human of emotions, that really makes this novel truly intelligent.

Advertisements

Botnik vs Harry Potter

hp01

“I’m Harry Potter!” Harry began yelling. “The dark arts better be worried, oh boy!”

When we last wrote about computer algorithms producing works of creative writing, we were talking slightly off-kilter poetry from the ‘mind’ of a program called OGDEN. Now we’re back on the topic again – only this time we’ve abandoned poetry in favour of Harry Potter; the greatest selling book franchise of all time.

Less an advanced computer algorithm and more a simple predictive text keyboard, Botnik describes itself as “a community of writers, artists and developers collaborating with machines to create strange new things.”

In their latest project, the team behind Botnik fed the machine the entire volume of seven Harry Potter novels, and then asked it to come up with a new chapter for the franchise.

The result really is quite something. Titled, Harry Potter and the Portrait of What Looked Like a Large Pile of Ash, you can read the latest chapter of Potter and co online (which you should do right now).

To give you a flavour of what’s in store, check out the first two pages below:

hp02

hp03

What seems so glorious about this endeavour is the feeling that for all its clear absurdity – “They looked at the door, screaming about how closed it was and asking it to be replaced with a small orb. The password was ‘BEEF WOMEN,’ Hermione cried.” – there is still some semblance of Harry Potter and J.K. Rowling to it all.

While we are a little in awe at the surreal genius of some of the lines Botnik has created, there is nothing fancy about these machines. They are not magically complex. They are simple algorithms built by simple tools. They follow predefined rules of grammar and structure to compose what they perceive as logical-sounding snippets.

The passages do reveal, however, interesting patterns within the lexicon of the Harry Potter franchise. The lines that do make sense, or perhaps don’t feel out of place – “Leathery sheets of rain lashed at Harry’s ghost as he walked across the grounds towards the castle.” – come across this way precisely because as readers we are used to seeing this type of descriptive exposition put down in this type of order. The words Botnik sometimes chooses may not always fit the bill – “he immediately began to eat Hermione’s family” – but they are presented (for the most part) in an order and structure that J.K. Rowling utilises most of all. In this way, Botnik holds up a fascinatingly surreal mirror to the writing voice and style of one of the best selling authors in history.

So, it would be thoroughly fascinating to find out what J.K. Rowling herself made of this quasi-A.I project. Does she see herself in some of the passages of Botnik-Potter? Or, perhaps the more intriguing question focuses on the machine in all of this – and so we might better ask Botnik whether it dreams of Electric Harry Potter.

Rise of the machines: will computers replace human beings in the publishing industry?

 

Hal_9000_in_2013_by_athdarariddle-d67149m

We’ve previously written about the way data on reader’s habits stands to revolutionise processes in the publishing industry. But while so-called “big brother analytics” might change the way publishing houses choose which books they invest in, a general assumption was that the ultimate decision would be made by a human being. This might sound overly obvious; but a recent development could potentially change all that.

In fact, we may be moving toward a world in which computers – rather than human beings – have the final say as to which books are published, and which books companies invest the heaviest amount of marketing funds in.

This all hinges on the success of a new project by data-driven publisher, Inkitt, in collaboration with Tor Books. And the two companies are now set to release the first novel selected by a computer algorithm for publishing.

Bright Star, a young adult novel by Erin Swan, was discovered using predictive data that analysed reading patterns on the Inkitt platform.

“This book deal sends a clear signal to the publishing industry that predictive data analysis is the way of the future,” says Inkitt’s Founder and CEO, Ali Albazaz. “Inkitt is at the forefront of the movement to use predictive data in publishing and this deal shows that our business model works. We are so excited to be able to help Erin kick off her career as a novelist – and we already can’t wait to get our hands on the next book in the Sky Rider series.”

Self-described as “the Hipster’s Library”, Inkitt functions as a platform that allows users to read books that haven’t been published yet – or to “fall in love with novels before they go mainstream.”

While some may point out that there isn’t very much hipster-esque about a company that has to tell people how hipster it is, what is interesting is how these developments may change traditional publishing models. Indeed, could this spell the end for the standard process of a qualified literary editor reading through manuscripts and deciding to invest in those they believe are the best fit for both their company, and for the wider literary industry?

Well, perhaps there is reason to believe so. Some of the most commercially successful novels – think Harry Potter or Twilight – were ignored by a succession of mainstream publishing houses before being picked up by organisations that ultimately reaped huge financial rewards for doing so. Using an algorithm to test what works best with readers could – in theory – help reduce the chances of a publishing house missing out on the opportunity to publish these sorts of best sellers.

But there are of course many caveats here. Not least of which is the fact that we are yet to see how successful Bright Star will be. But furthermore, we may also wish to question whether we truly want a publishing industry built upon the decisions of machines.

It’s true that other algorithms have been designed to make it appear as though computers can write poetry (and some of these AI poems have even been published). Yet there is something innately human about literature and writing. And with books occupying such an important part of our culture, it does seem a risk to remove the human being from the equation.

A further risk here, of course, is that an algorithm designed to identify books that have the greatest financial value in them may not actually be the best books. Fifty Shades of Grey may be taken as an example here – for it stands as an example of a trilogy of books that have sold tens of millions of copies, despite the writing being of questionable quality. These are the books, after all, described variously as “stilted and cliché-ridden” (New York Review of Books), “reading as though women never got the vote” (the London Review of Books) and even as “extremely dangerous […] [because] the themes of the novel – love alone can make someone change, that abuse from a spouse is acceptable as long as he’s great in bed, that pregnancies should always be carried to term even if the parents are not ready to be parents, and the ridiculously antiquated, Victorian idea that the pure love of a virgin can save a wayward man from himself – are irrational, unbelievable and dangerous”.

What are the risks that, should the publishing industry come to rely on computers to make decisions – rather than experienced editors and industry professionals – we come to develop a cultural void in which every book is published not for its merit, but because of its ability to sell copies? What are the risks that we create a cultural imbalance within literature, where our literary canon is filled of, essentially, thousands upon thousands of books like Fifty Shades of Grey?

This is not to disparage readers of the E.L. James novels – but to argue that our culture relies on variety, rather than similarity. The great thing about books is surely that they can cater to all tastes – and anybody can find familiarity and connection with some book, somewhere. And it seems that an industry run by machines motivated purely by the pursuit of commercial success can only serve to narrow the selection of books available to us.

There are already signs that this is taking place already. As pointed out in this Litro Magazine article points out, “there is an increasing focus on mimicking commercial success, rather than striving to create something that is new.” And the influence of modern neoliberal capitalism has seen the publishing industry gradually follow the film and music industries in only investing in pieces of art that seem geared towards bringing in money, rather than new ideas. As such, the industry is increasingly dominated by novels that are copies of novels, which are themselves copies of other commercially successful novels.

In fifty years, will Inkitt and its publishing algorithms be regarded simply as a minor curiosity? Or part of the start of an AI revolution within human culture? If it’s the latter, we may have just witnessed what will come to ultimately eliminate and replace human beings from publishing. Ultimately, it’s up to you, dear readers, how you feel about that.

Let us know your thoughts in the comments below!

And, as always, remember to sign up to our free newsletter, so you never miss a trick. Join the gang!