This is a collection of short essays summarizing in parallel format selected world capital city histories, as its title declares. Its 15 city narratives are remarkably consistent, keeping to a narrow definition of administrative history – although one might criticize a few more impressionistic contributions (among which I would disappointedly number London) and also a few swamped by the complexity of their subject (among which would come Brussels). Editing is irregular (some incorrect English and computer glitches – Brussels again), but length and focus achieve admirable uniformity. Here lies, for me, much of the value of the book: it presents a consistent body of comparative material and I would use it as an introductory text for a city design course, around which to arrange more critical and problematic readings to explore the considerable tensions showing through Gordon's tight grid.

Introductory essays by editor Gordon and Peter Hall – both planning historians – plus a concluding essay by Hall make it clear that their interest is typology: what distinguishes capital cities? What is their place among the emerging mega-cities of today? How is their function as capitals balanced with their nature as cities – how here are the representational and the economic balanced? This last issue is brought out by a third introductory chapter by the architect/planner Lawrence Vale on the design nature of capitals, working off the common assumption that a capital should be recognizable as such – like Washington DC and unlike (again) Brussels. Yet for Gordon and Hall size is the distinguishing factor as they insist on placing the cities among lists of the largest agglomerations on the planet in constructions familiar from econometrics and the books of Saksia Sassen. They squeeze in a last essay on New York largely on these grounds.

Not all capitals are big, and if small cities are included, why Ottawa and Helsinki rather than Pretoria and Rabat? (Africa is completely missing from the book.) If one provincial capital is included, Chandigarh, what about others like Le Corbusier's edgy Algiers projects or Nelson Rockefeller's Albany? What about war capitals: Changchun in Japanese Manchuria, Strasbourg under German control (1871–1918, 1941–44). Or purely symbolic capitals like Hendrik Andersen's 1913 Centre mondial manifest as Ernest Hébrard's lush Beaux-Arts drawings – the quintessence of the capital type, all the more so for Andersen's suggestion that it might be constructed anywhere on earth (providing maps to show just how).

How might one organize a survey like this one to give it cohesion and force? Wolfgang Braunfels (1976) defined city types (Cathedral Cities, City-States, Princely Seats . . . Capital Cities); Thomas Hall (1997) sped through a string of nineteenth-century European capitals as here, but sandwiched this between a general introductory chronology and a series of shared issues (Motivation for Plans, Authors, Decision Processes, Content and Purpose, Elements of Plans . . .); Wolfgang Sonne (2003) focused on a group of examples around 1900 to explore the architectural idea of the capital Gestalt. Donald Olsen (1986) made issues – expanded to broad, cultural ones – the centre of his study, using only three European capitals (London, Paris, Vienna); Carl Schorske (1980) famously used only one example, Vienna, but explored cultural character with tremendous precision. But in the case of capitals, things seem inevitably to topple back into narratives of plans broached, drawn, debated and set aside as in Hall's chapter on decision making, Arturo Almandoz's edited Planning Latin America's Capital Cities (2002) and now in David Gordon's book under review here.

This emphasis on typology and administrative organism deflects discussion from several basic and illuminating tensions making themselves felt as one moves from essay to essay: 1. The steadily increasing isolation of capital complexes within the immensity of most twenty-first-century agglomerations like New Delhi; 2. The counter-model of capitals with representational buildings scattered across their texture, as in London, Paris and especially Brussels; 3. The un-naturalness of fixed representational complexes to begin with. At the heart of these issues may be a fact that this book avoids: centres of power in recent history have been remarkably evanescent. For example, in South Africa since union in 1908, the legislature meets in Cape Town, the judiciary in Bloemfontein, the administration sits in Pretoria (in a complex resembling a capital – which it isn't – with a counter-symbol, the Afrikaner Voortrekker Monument, facing it across the valley). Brussels is really more a figure of speech for the seat of the European Community, with Strasbourg, Luxemburg and Frankfurt functioning parallel to it while it remains capital of Belgium.

Perhaps this imprecision is the whole point: maybe capitals are fundamentally unstable as an idea and as a reality in 2010. This might be why corporate headquarters buildings like the UN or Berlaymont in Brussels can serve perfectly well as seats of power – there are many parallel sorts of power today in different sorts of clothing. The quaintness of the New Delhi ‘bungalow district’ protecting Lutyens’ and Baker's marble confections makes one aware of the wall of skyscrapers ever thickening around the Paris Boulevard Péripherique signalling its Haussmannian core as a huge symbolic indulgence. Perhaps what we need is a blunt analysis of the capital illusion at its roots. Brussels beckons – it might be time to follow.