Comment: Forum physicsworld.com pulls strings

In trying to interpret the “landscape” and the meaning of space– in a quantum theory of gravity, physicists should consider knocking on the door of their nearest philosophy department, Christian Darkin/Science Photo Library says Steven Weinstein

String theory, like any mathematical theory, presents physicists with a host of technical challenges in how to extract predictions about the real world. As a unified theory of gravity and particle , string theory should, on the face of it, be amenable to ex- perimental test because it must at the very least reproduce all observed phenomena. But extracting unambiguous predictions has proved difficult. Bound up with these ostensibly technical Different worlds Issues such as the nature of the can bring physicists and philosophers together. challenges are some rather philosophical conundrums. For example, how should the of predictive power has been to invoke the ask). Advocates of this anthropic interpre- theory be tested in view of the very many dif- . Originally employed by tation of the landscape – the most prom- ferent universes that it is able to describe, cosmologists to explain what appeared to be inent of whom have been Andrei Linde and and what is the meaning of space and time? improbable values of certain cosmological Leonard Susskind at Stanford University While fruitful collaborations between phy- parameters, the anthropic principle states and Alex Vilenkin at Tufts University, all sicists and philosophers are rare, there is that what we can expect to observe must be in the US – therefore suppose that it is little doubt that string theory forces us to restricted by the conditions necessary for our reasonable to expect that humans are such tackle issues that cross both disciplines. presence as observers. In other words, the typical observers. world is observed to be the way it is because Philosopher Nick Bostrom of Oxford Uni- The anthropic landscape that is the only way that we humans could versity in the UK, who calls this the “self- The first set of philosophically flavoured co- ever be here to ponder such questions in the sampling assumption”, has been one of the nundrums concern methodology. Although first place. more audible advocates of anthropic reason- string theory requires 10 (in some cases 11) The use of anthropic reasoning in the con- ing among philosophers (2002 J. Philosophy dimensions, it prides itself on having only text of the string-theory landscape depends 99 12). The present author, on the other one free parameter: the string tension. Ori- on a somewhat hand-waving extension of the hand, has argued against this style of reason- ginally it was hoped that the Standard Model currently popular cosmological model called ing by pointing out that the assumption that of particle physics, which has 19 free param- “eternal inflation”. In this model, “the” uni- we are typical observers is both ambiguous eters, would somehow fall out of string the- verse is simply “our” universe – one of an and unmotivated (2006 Class. Quantum ory as the unique low-energy, 4D vacuum infinite number of universes that make up Grav. 23 4231–4236). Meanwhile, cosmolo- state. As such, physicists would be able to the “multiverse”. The string-theoretic exten- gists Anthony Aguirre of the University of explain why these 19 parameters take the sion of this model posits that each universe California at Santa Cruz and Max Tegmark values that they do. may have a different low-energy vacuum of the Massachusetts Institute of Technology In the last few years, however, string the- state (i.e. it may correspond to a different have suggested that nature itself might give orists have discovered that there is a vast location in the landscape). String theory us a clue as to which sort of reasoning we “landscape” of at least 10500 such vacua, and would, it is supposed, ultimately assign a should employ (2005 J. Cosmo. Astropart. that each of these corresponds to a universe probability to each of these vacua – although Phys. JCAP01(2005) 013). with a different set of fundamental particles, given that researchers do not yet have suf- Proponents of anthropic reasoning tend interactions and parameters. Although the ficient knowledge of the dynamics to be able to see the huge number of possible vacua as existence of so many vacua makes it seem to derive such a probability measure, this is necessarily implying that a universe like ours “likely”, in some sense, that the Standard highly speculative. is inevitably a low-probability (and Model will show up in one of them, it now The hope of anthropic proponents is that therefore merely allowed by, rather than seems decidedly non-generic. Only if there the world we see is explained by its being predicted by, the theory). Other string the- were an infinite number of vacua would the typical among those universes that allow the orists, however, say that were somebody to situation be any worse from the standpoint existence of observers, intelligent life, con- discover that our universe was well described of uniqueness. scious life or something along these lines by a statistically unlikely vacuum, string the- One reaction of string theorists to this loss (preferences differ depending on who you ory would nevertheless be judged to be

18 Physics World September 2007 physicsworld.com Comment: Forum

perfectly adequate. Discomfort with the low The future(s) probability might lead us to question the as- Quantum theory Finally, philosophy could help with the quite sumptions that gave rise to our probability general question of whether there are ways distribution in the first place, as string the- could be an artefact to describe the world that do not involve mat- orist Michael Douglas of Rutgers University ter being distributed in three large (i.e. non- has suggested. Or we might simply point out compactified) space dimensions and one that no matter what the distribution is, any of the assumption large time dimension. String theory and other particular vacuum is highly unlikely, and so approaches to quantum gravity – indeed all we should be no more surprised at observing that the world contemporary physical theories – assume one rather than another. that our world is well described on ordinary The general problem of assigning a prob- energy and length scales as a 4D space–time, ability distribution to a set of possibilities in has a single and go on to contemplate elaborations (extra the absence of any prior knowledge has been spatial dimensions in the case of string the- discussed by philosophers over the years, time dimension ory) and modifications (such as quantization especially in conjunction with attempts to of space–time in the case of loop quantum ground inductive reasoning in Bayesian observations are usually idealized as taking gravity) of this basic structure. But long- methods. Indeed, John Norton of the Uni- place in local regions of classical space–time. standing difficulties in interpreting quantum versity of Pittsburgh has recently argued that Once one addresses such questions about mechanics, such as the measurement prob- one can, and sometimes must, characterize what it means to formulate physics in the ab- lem and the emergence of classicality, and the complete ignorance about a physical system sence of a classical space–time background, conceptual and technical difficulties involved without appealing to a probability distribu- one is then confronted with the question in non-perturbative quantum gravity could tion at all (Philosophy of Science at press). In of what it means for classical space–time mean that this picture is incorrect. fact, Douglas and Shamit Kachru of Stan- to “emerge” from the theory. In ordinary For instance, one might contemplate mul- ford have recently suggested something quantum mechanics, some physicists and tiple time dimensions. Physicists Itzhak Bars along these lines as a way of characterizing philosophers think it legitimate to postulate of the University of Southern California the landscape (Rev. Mod. Phys. 79 733–796). that classical behaviour arises via the inter- (2001 Class. Quant. Grav. 18 3113–3130) and action of quantum systems with a macro- Chris Hull of Imperial College London Tackling space–time scopic environment, while others object to (arXiv:hep-th/9911080) have explored dif- When dealing with any theory of quantum this sort of “open system” analysis and in- ferent ways that multiple might enter gravity – including, but not limited to, string stead characterize the emergence of classi- into string theory. Another possibility is that theory – one must eventually face deep con- cality as being merely apparent. This, for quantum theory itself is an artefact of the ceptual issues that relate to the quantization example, is the view of “consistent histories assumption that the world has a single time of space–time. String theory has for the most programmes”, which attempt to explicitly dimension. For instance, perhaps a “classi- part postponed dealing with these issues, as realize the many-worlds interpretation of cal” theory with multiple time dimensions the theory itself currently consists of an inter- quantum mechanics. could explain non-local phenomena such related “web” of perturbative theories that In both cases one ordinarily postulates a as entanglement. makes sense only at energies low enough background space–time. In the open-system This idea is highly speculative, of course. that space–time is effectively classical. analysis, the interaction with the environ- But explorations along these lines would However, at sufficiently high energies one ment is a process that occurs over time; while seem to be an excellent collaborative exer- expects that the classical approximation of in the consistent-histories approach, the cise for philosophers and physicists, given space–time will break down, and instead the background space and time characterize the the closely intertwined nature of the philo- non-perturbative (and unknown) theory quasi-classical histories that are supposed sophical and mathematical aspects of the at the centre of the web – sometimes called to emerge from the quantum theory. In a full problem. We can of course resist such spec- M-theory – must be invoked. In such cir- theory of quantum gravity, however, one ulation by simply insisting – in the spirit of cumstances one must face squarely the prob- expects there to be no background space– Kant – that it is an a priori truth that ob- lems encountered already by researchers time at all, and so a new notion of emergence servers are properly idealized as time-like working on alternative theories of quantum is required. world lines in a 4D space–time. But we gravity, such as loop quantum gravity and In canonical approaches to quantum grav- might be more adventurous, in the service other “canonical” approaches. ity these difficulties have been addressed, of both physics and philosophy. Among these is the notorious “problem of if not resolved, for example in the work of time”, which concerns how one should for- philosopher Jerermy Butterfield of Cam- Steven Weinstein is in the mulate a quantum theory in which there is no bridge University and physicist Chris Isham Philosophy Department at the classical spatial and temporal structure with of Imperial College London (arXiv:gr-qc/ University of Waterloo and respect to which one can characterize phys- 9901024). In string theory, however, the is an affiliate of the Perimeter ical states and their evolution with time. notion of emergence begs for further dis- Institute, both in Canada, There is also a related problem of how to de- cussion, and could lead to fruitful collabor- e-mail [email protected] fine observables in such situations, given that ation between physicists and philosophers.

What Matters in Materials Research? A Technical Publication from Sigma-Aldrich Materials Science TM For a complimentary subscription to Material Matters Sigma-Aldrich Materials Science—Your Chemistry Partner visit sigma-aldrich.com/mm3. sigma-aldrich.com

Physics World September 2007 19 02288PhysicsWorldad.indd 1 8/14/2007 7:59:30 AM