This anthology—published last year to mark the first Japanese Worldcon—is intimidatingly generous. Before you even get to the fiction, there are forty-one pages of stuff: a biographical sketch of Judith Merril, which makes clear the debt this project owes her; a short preface by David Brin, which doesn't say much; an actual introduction ("Phase Shifting") by one of the editors, Gene van Troyer, which provides a crash course in the history of Japanese science fiction and fantasy in translation (following Merril's preference they use "speculative fiction," as will I in this review); and a note by Tatsumi Takayuki introducing a critical essay by Shibano Takumi (1971, revised 2000 and previously published in Science Fiction Studies #88). Then there are fourteen stories and one poem (originally written between 1962 and 2002), followed by a brief historical sketch by Asakura Hisashi, an equally brief paean to translators by the other editor, Grania Davis, and thirty-one biographical sketches covering everyone who's contributed to the book in any way (twenty-five men, six women): that is to say all the authors, all the translators, the editors, the cover artist, and the office dog (okay, not that last one). In all, something over a sixth of what is labelled "outstanding tales of Japanese science fiction and fantasy" is context. It's all entirely welcome and not a little overwhelming.

So it is awfully tempting to take the book at face value and treat it as a representative sample, a literary tradition in a thimble. But a few caveats need to be attached. First: over half the stories included were first published in the 1970s, and most of the rest were published in the 1960s. There are only two stories from the eighties, none from the nineties, and one from the present decade. Given that the first Japanese science fiction magazine to manage more than one issue (with the original title of SF Magazine) appeared at the end of 1959, slightly preceded by the first fanzine, Shibano Takumi's Uchujin (Cosmic Dust), it's clear that this is predominantly a sampling of Japanese speculative fiction's early and middle days. The second caveat goes some way to explaining this skew: as hinted above, this anthology represents the culmination of a project started by Judith Merril in the mid-seventies, and in fact includes five stories that appeared in the only previous major book anthology, The Best Japanese Science Fiction Stories (1989, eds. Martin Greenberg and John Apostolou), plus a number of stories that were only partially translated at that time. My third caveat is that even a brief survey of the author biographies suggests that the borders between different media may be more fluid than they are here—a substantial number report working on manga or anime, alongside prose projects—which may also affect the portrait painted. (Mind you, apparently two of the authors are also horse racing commentators, and I'm not sure what you'd want to read into that.) For completeness, I'd also note that only two stories are by women—the two most recent, in fact, although I have no idea whether that can be attributed to anything more than chance—and that almost all of the stories were published when their authors were in their thirties or early forties.

Van Troyer's introduction glosses the book's contents as "an idiosyncratic cross section of the history of Japanese SF over the last fifty years"; nevertheless, he says, it covers the "three major phases" in the development of the genre, as identified by Yamano Koichi in his 1969 essay "Japanese SF, its Originality and Orientation" (published in English in 1994). Van Troyer doesn't get into the essay's historical context, but it's worth pointing out—as Darko Suvin's introduction to the essay's SFS publication does—that it was written before Japan achieved the economic success familiar to us today. It's probably also worth noting Suvin's claim that Yamano was writing "against the grain of dominant trends in Japanese SF, as concerns both the writers and the readers' taste." Nevertheless, the three phases are broadly what you might expect: a first phase ("The pre-fabricated house phase," or "infiltration and diffusion"), in which Japanese writers are most strongly influenced by Western traditions; a second phase ("Remodeling the ready-built home phase," or "adaptation and acquisition"), in which the scope of Japanese writers expanded; and a third phase ("Putting up a new house," or "creative departure"), in which Japanese speculative fiction develops its own cultural identity—although what exactly that identity might be is never clearly articulated. Yamano theorises that it would be comparable to the position of scientific romance as a distinctively British style of science fiction, which is one of those statements that seems clear until you start thinking about it too hard, particularly given that there are a good number of pure fantasy stories included. It's also not possible to really assess how accurate this model is on the basis of one collection, but it is interesting that although the two examples of third-phase Japanese sf are the two most recently published stories, the seven examples of first-phase Japanese sf were first published between 1962 and 1981 (twelve years after Yamano's essay!), while the five second-phase stories appeared between 1963 and 1974. It may be that the three phases can be thought of as three coexisting styles; and they do make a convenient way to organize a review.

Before getting into the details of the fiction, it's worth one more digression to provide a reminder of the obvious: a work in translation is not the same work as the original. As Davis's afterword makes clear, the challenge of idiomatic translation is as great between Japanese and English as between almost any two languages you care to name; so inferring anything about differences in literary sophistication or maturity (as a three-phase model encourages us to do) from work translated by many hands, as the stories in this collection are, is to step out on thin ice. But I think it's safe to say that there are more-sophisticated and less-sophisticated handlings of genre materials in this book—in the sense that your typical noughties Asimov's story is more sophisticated than your typical thirties Astounding story—and it does seem to be the case that the two third-phase stories are at the top of the tree in that regard.

Ohara Mariko's "Girl" (1985, translated by Alfred Birnbaum) builds a vivid setting—a decadent, far-future City on the verge of collapse—and tells a complex and moving story in relatively few pages. The protagonist, Gil, is an erotic dancer whose upper body has been resculpted to appear voluptuously female, but whose lower half remains resolutely male. An encounter with a more traditionally gender-ambiguous girl ("The girl looked angelic in profile. But weren't angels boys? A girl, the image of a boy," p. 156) leaves Gil intoxicated, nearly hypnotized. Gil's neuroses and sense of isolation are mercilessly anatomized, but at the end the camera lens pulls back, and for all the looming darkness the story ends on a note of purest grace, Gil's story reframed as "a light made visible by the dark" (p. 169). To throw around too many comparisons to familiar Western writers would be overly reductive—these stories, and writers, should be allowed to stand for themselves—but I'm sure I won't be able to resist here and there, and I do think there's something in the intensity of "Girl" that recalls the best of James Tiptree, Jr. (Not to mention the subject matter, which stands out starkly next to some of the stories discussed below.) The other third-phase story, Kawakami Hiromi's "Mogera Wogura" (2002, translated by Michael Emmerich), is less intense, but equally concerned with identity. Structured as a prototypical day-in-the-life (morning, lunch, afternoon, and so on), the initial hook is the need to piece together exactly what manner of weirdness is going on. At first, all we know is that the narrator has a number of humans living in his house; how they got there, for what purpose they were taken, and exactly what sort of creature the narrator is are initially all open questions. Gradually we come to understand that he (this may be a prejudicial assignment on my part: the narrator has a wife, but is never explicitly confirmed as male) is some sort of humanoid (shortish, hairy, clawed), whose people have coexisted alongside humanity for centuries (as opposed to, say, an extraterrestrial). In a nice touch, in the story's world the existence of these creatures is no longer secret, and in fact it transpires that the narrator works in an office, where he identifies the humans he will "collect." The reason he collects them has something to do with sympathy, and something to do with curiosity, and perhaps even simply something to do with habit; "Mogera Wogura" is a story confident enough to present its situation calmly, knowing that it will fascinate even if it doesn't provide all the answers.

In contrast, many of the stories identified by van Troyer as first-phase share a familiar trait of Golden Age science fiction stories in English, that of being keen to share their answers. The earliest and latest—Mayumura Taku's "I'll Get Rid of Your Discontent" (1962, translated by M. Hattori and Grania Davis) and Kajio Shinji's "Reiko's Universe Box" (1981, translated by Toyoda Takashi and Gene van Troyer)—are both gadget stories. The former is science fiction couched as fairy tale, in which a salaryman acquires a strange object—described as "a short blueish plastic won-ton placed in a small plastic box" (p. 87)—from a "shabby shop." The instructions promise that, if grasped, the object will relieve his feelings of anger, anxiety, or embarrassment, but caution that it can only be used three times, and that if it is used a fourth time "something fatal will happen to you" (p. 87). It's an ambiguous phrasing that Mayumura exploits as well as he can, given the story's short length, but ultimately the story is all about its punch line. Kajio's story features the most explicit dialogue with Western science fiction in the book, in that its conceptual debt is acknowledged by having the titular box, which works exactly as advertised, delivered to a pair of newlyweds from "Fessenden & Co.," this being a nod—noted by Brian Stableford in his review for The New York Review of Science Fiction—to Edmond Hamilton's story "Fessenden's Worlds" (1937). The story unfolds pretty much as you'd expect: one member of the couple (the husband) finds the box dull and irrelevant, while the other (Reiko, his wife) becomes fascinated with it; and the climax of the story involves a conjunction of box and relationship. Perhaps the most interesting aspect of the story is the rigid demarcation of gender roles, in that the reason Reiko has time to become fascinated with the box is that, once married, she stays at home while her husband Ikutaro goes out to work, feeling that he must work harder and longer than his colleagues in order to provide the best for her. That this is an unstable system isn't surprising, but the strength with which it constrains the characters is something to behold, and the contrast with the more fluid systems in "Girl" is striking.

Other first-phase stories are also variations on forms familiar to us, or, to use Yamano's phrase, the "creative methods" of Western science fiction. Ishikawa Takashi's "The Road to the Sea" (1970, translated by Yano Tetsu and Judith Merril) is about a boy on a journey who discovers an essential truth about the world he lives in. The secret is successfully obscured until the reveal, although I was distracted by some of the translation: "No slings or arrows," we are told, would stop this boy on his journey, and it's hard not to wonder whether the reference to Shakespeare is in the original. In "The Flower's Life Is Short" (1967, also translated by Yano and Merril), by Fukushima Masami, the background is more interesting than the foreground. The story extrapolates the future of ikebana (Japanese flower arranging), as well as a couple of social changes—extreme longevity has been achieved, and marriages are categorized according to whether they are short-term or permanent contracts. These developments are conveyed with admirable economy, if not elegance, and some of the descriptions of electronic flower displays are lovely, but the character dilemma at the heart of the story—an artist confronting her loneliness—feels merely contrived and melodramatic. More interesting is Toyota Aritsune's "Another Prince of Wales" (1970, translated by David Aylward), set in a world in which unsupervised warfare has been recognised as simply irresponsible, so all conflicts are arbitrated by the UN (and watched on TV by millions of obsessed fans: soldiering is the easiest way to global celebrity). The story's narrator, Keith Kimura, is a half-English, half-Japanese "UN War Supervisor," and it's noticeable that Toyota feels obliged to hang a lantern on his heritage: "these days, a racially mixed child like me is quite common, and even a mixture of three or four races like [his colleague] Isabelle is not rare" (p. 126). Inevitably, the story portrays a war between Japan and England (not Britain), and equally inevitably this draws out the conflict we are meant to believe is inherent in Kimura's heritage. Toyota is no Kazuo Ishiguro: for him, the differences between Japanese and English nature are more important than the similarities.

The remaining two stories that van Troyer labels as first-phase seem to me to be closer to the anthology's second-phase stories in complexity, concerns, and tone. Yano Tetsu—one of the "Uchujin coterie," and a key figure in Japanese speculative fiction up until his death in 2004—is represented by fiction as well as his translation work. "The Legend of the Paper Spaceship" (1975, translated by Gene van Troyer and Oshiro Tomoko) is the longest story in the book, couched as the experiences of a soldier dispatched to a remote mountain village during World War II. Initially, he seems most interested in Osen, the village madwoman or idiot (depending on who you ask) and, for the men of the village, their whore; later, the focus is on Osen's son Emon, who is psychic in a way not unlike Arun in Vandana Singh's recent novella Of Love and Other Monsters (2007), and to not dissimilar effect; towards the end, the narrator admits he isn't sure which of them is the true protagonist. The answer seems to reside in the origins of the village, which is strangely isolated ("an island in the stream of history," p. 213), and whose folk songs and memories (as accessed by Emon) may hide a secret of the kind hinted at by the story's title. There's much to like about the story, not least the confidence with which Osen and Emon are presented as individuals rather than victims, but it is hobbled by some unavoidable translation awkwardness: specifically, the narrator's attempts to decode the aforementioned songs, which involves laboriously explaining homonyms and mistranslations that (one assumes) must flow more smoothly in the original.

No such infelicities mar the anthology's commanding opening story, Komatsu Sakyo's "The Savage Mouth" (1969, translated by Yano and Merril). Komatsu, Worldcon Guest of Honour, is apparently Japan's equivalent of Isaac Asimov, but I'm pretty sure Asimov never wrote anything quite as visceral as this tale of literal self-consumption. The nameless narrator of "The Savage Mouth" is a man who seethes with anger at the absurdity of life, so decides to turn his back on the world and, piece by piece, replace his body with cybernetic parts and eat the flesh so liberated. He stews his shin; uses his duodenum to make sausages; eats his stomach soaked in garlic and pepper. This continues to a logical conclusion, whereupon the viewpoint shifts to a weary police inspector, reviewing the scene and declaring that the people at large must never know because they "would begin to lose confidence in their own behaviour" and "start peering and probing at the blackness inside their own souls" (p. 51), thus trumping the early body horror with a deeper and more lasting metaphysical horror. It is the best of the first-phase stories; although as I suggested above, for its examination of the line between human and machine—not in the first rank of the themes of Western SF—I think "The Savage Mouth" could be as comfortably described as a second-phase story. Indeed, in his essay Yamano identifies Komatsu as a second-phase author, although he doesn't discuss "The Savage Mouth."

Even excluding Komatsu's story and "The Legend of the Paper Spaceship," however, the five second-phase stories are a much more consistently satisfying clutch than the phase-one offerings. The weakest is probably Hirai Kazumasa's "A Time for Revolution" (1963, translated by David Aylward), which has a marvellously chill opening—"They stood huddled at the bottom of the atmosphere, shivering in the cold" (p. 53)—but quickly retreats to a surprisingly generic near-future city and a surprisingly Luddite resolution. It's of a piece with the other phase-two stories, though—all of which, interestingly, were translated by Dana Lewis—in its fascination with identity and transformations. In some cases this is played for laughs: Hanmura Ryo's "Cardboard Box" (1975) tells the exceedingly literal and impressively lewd story of, yes, a cardboard box. Emptiness, you see, is a form of torture, while being filled is very satisfying. More often, however, transformation is something eerie and of uncertain benefit. The narrator of Kono Tensei's "Hikari" (1976)—which means "light"—is taking the night train and sees, in the distance, a shining city of light. His companion offers to explain what happened there, and in best club-story fashion tells a tale about a strange family with bright eyes who moved into his town and started cleaning it, night and day. Gradually they converted most of his neighbours, even his wife and children, to their crusade and ultimate apotheosis; and, it seems, they have visited several towns in this area, to the same effect each time. "Hikari" is an odd but haunting story, deeply sceptical of transcendence; as in "A Time for Revolution," "Mogera Wogura," and a couple of others, the characters are largely nameless, which reinforces the dreamlike tone—although it's still more obviously science fictional than, say, most of Murakami Haruki's work. You suspect Kono would agree with Tsutsui Yasutaka, who, according to his biography, thinks of speculative fiction as "an approach to deconstruct reality," and whose "Standing Woman" (1974) is the closest the anthology comes to Sterling-slipstream. As in "The Flower's Life Is Short," there are two innovations, one technological and one social: the latter is that society has become more totalitarian, such that it's illegal for two or more writers to assemble together, and the former is hybridisation of various kinds of animals with plants. The urban megalopolis in which the narrator lives is dotted with such hybrids, known as pillars—dogpillars, catpillars, manpillars—which can be instituted as memorial or as punishment. The story is convincingly strange and alienating, from the descriptions of the converted creatures (catpillars: "greenish faces stiffly set," p. 175) to the restrained pain of the narrator and his wife.

Which just leaves Yamano Koichi himself. Apparently, Judith Merril described Yamano as "the Japanese Michael Moorcock," and certainly his exquisite story, "Where do the Birds Fly Now?" (1971) has something of a New Worlds air to it. (From his SFS article, his drubbing of US writers such as Heinlein ["banal realism"] and Harry Harrison [on Deathworld: "I must wonder whether the writer was in this story actually serious about creating a consistent novel"] is also entertainingly feisty.) The narrator—nameless again—relates his experience with an unusual species of bird, described as flying "as though striking a canvas boldly with a brush" (p. 103), that has somehow evolved to fly in four dimensions. The interaction between the birds' flight and the narrator's mind drags him across the walls between worlds, and the various elements of his life—his girlfriend, his involvement with the local antiwar movement, the old man who explains the migratory patterns of the birds—reconfigure around him as he travels, starting with a shoot-out that seems (the next day) never to have happened. Most impressively, Yamano structures his story into a number of sections, of which the first and last are designed to be read first and last, but which can otherwise be read in any order the reader wishes, providing a pleasing series of dislocations and epiphanies. The final sections intensify both the emotional and speculative elements of the narrative; and at the end, the narrator has travelled further—travelled in mind as much as in body—than we might have expected. In his introduction, van Troyer notes that second-phase stories "often distanced themselves from traditional Japanese cultural perspectives, foregrounding a Western-style 'rationalistic' and objectively macroscopic world-view" (p. 10); I can't think of a better way to praise "Where do the Birds Fly Now?" than to say it couldn't be further from that description.

But there is something about the anthology that defies expectations. Yamano's model seems relatively sound—at least as a way to understand the stories between this particular set of covers—but of course, written forty years ago, it doesn't encompass what most people (I would guess) think of when you say "Japanese SF" to them. Reviewing a debut anime (by Fumihiko Sori) in Interzone 217, Nick Lowe notes that it "deals sweepingly in the great Japanese themes: the corporatisation of society and industry, and the destruction of the individual; technological supremacy weighed against environmental guilt; the politics of isolationism and the preservation of cultural identity against Western intrusion; and above all, a climax where Toyko gets totally Gojiraed by giant sandworms of pure mecha." It's noticeable that the stories in Speculative Japan largely stay away from these themes, or touch on them in attenuated form, and focus instead on those issues noted above—personal identity and metamorphosis, most prominently. Yet Lowe's list doesn't strike me as a desperately unfair assessment of what's distinctive about the Japanese speculative fiction that's reached these shores. Obviously most of it postdates most of the work in this book ("Girl" does have more in common with familiar anime settings than any of the other stories); equally obviously the difference may be a reflection of a prose/anime divide (if all that had made it to Japan of the Western genre was Star Trek and The X-Files they'd have a skewed picture of our field, too), or simply differing taste on the parts of the relevant gatekeepers. Whichever, it's a reminder that whatever it is, Speculative Japan isn't a tradition in a thimble: it's a keyhole. It would be good to open the door.

Niall Harrison (niall.harrison@gmail.com) has reviewed for The New York Review of Science Fiction, Foundation, and Bookslut, among other places. He blogs at Torque Control.