The Longest Journey: The Adventure Genre's Late Flowering

The timing is the first interesting thing about The Longest Journey. Funcom’s four-CD point-and-click adventure shipped in Norway as Den lengste reisen in late 1999 and reached English-speaking players the following year, which is to say it arrived at the exact moment the trade press had settled on the idea that nobody made these any more. Grim Fandango had come out in 1998 to reviews that were essentially eulogies. Sierra’s adventure line was being dismantled. LucasArts had one more in the tank and then stopped. Into that, a studio in Oslo delivered a fifteen-to-twenty-hour game built almost entirely out of talking.

It sold. Modest numbers by the standards of the shelf it sat on, and steady enough to spawn Dreamfall: The Longest Journey in 2006 and, after Ragnar Tørnquist left Funcom to found Red Thread Games, the crowdfunded Dreamfall Chapters across 2014 to 2016. That’s a twenty-year arc off the back of a genre everyone had agreed was over. Worth asking what it had that the obituaries missed.

Two worlds, one lever

The structure is the cleanest thing in it. April Ryan is an eighteen-year-old art student in the Venice district of Newport, a city in a world called Stark — recognisably ours, pushed forward a couple of centuries, all rain and neon and rent she can’t quite make. The other world is Arcadia: magic, guilds, dragons, a medieval-ish register. April is a Shifter, meaning she can move between them, and the game’s central lever is that she can do it more or less at will once she learns how.

That lever does an enormous amount of design work. Most adventures of the era gave you a set of rooms and an inventory and asked you to brute-force the intersection. The Longest Journey gives you two complete grammars and lets you carry objects and knowledge across the seam. A thing that is mundane in Stark is an artefact in Arcadia. A problem that is magical in Arcadia has an engineering answer in Stark. Once you internalise that, the game stops being about clicking on scenery and starts being about which world a given problem actually belongs to.

The writing knows this is the good idea and leans on it. Stark’s chapters are shot through with a specific late-nineties anxiety — surveillance, corporate housing, the sense that the technological world has been optimised into something joyless. Arcadia is warm and slightly ramshackle and full of people who want to talk to you. April is homesick for a place she’s never lived. The game earns that by making the two halves feel genuinely different to play, not only to look at.

The dialogue is the mechanic

The Longest Journey has a reputation for being talky, and the reputation is fair. April will describe a wall to you. She will have opinions about the wall. She has a full inner monologue and she runs it constantly, and if you click on everything in a room you can easily spend ten minutes there without solving anything.

I’d argue this is load-bearing rather than indulgent, and here’s the mechanical reason. In a genre where the standard failure state is I have no idea what the designer wants, April’s commentary is the hint system, the characterisation and the world-building running through a single channel. She tells you what she notices, which tells you what matters. She tells you why she won’t do something, which rules out a whole branch of your guessing. She’s funny about it, which means the hint doesn’t feel like a hint.

Compare that to the LucasArts house style, where the protagonist is a delivery mechanism for jokes and the hinting is done by object descriptions. Both work. Tørnquist’s version buys something extra: by the end you know April well enough that the plot’s demands on her land. The cost is pacing, and the game does sag — the middle chapters in particular ask for patience that a 2023 player, trained on a decade of games that respect the clock, may not have.

The duck, and what it tells you

You cannot write about this game without the duck. There is a puzzle where April needs to retrieve a key that has fallen where she can’t reach, and the solution involves an inflatable rubber duck and a clamp, assembled into an improvised grabber. It is the most-mocked puzzle in the game and it has been mocked for over twenty years.

Here’s the thing: it’s a bad puzzle for an interesting reason. The logic is sound once you see it. The failure is that nothing in the preceding hour teaches you that this game does improvised-tool physics. The rest of The Longest Journey solves problems socially — you talk to someone, you learn a fact, you use the fact. The duck asks you to switch grammars with no warning, and adventure players of that era responded the way they always did, by trying every item on every other item until something clicked.

That’s the genre’s structural disease, and it’s the real subject of why point-and-click died and what replaced it. A puzzle is a conversation about rules. If the rules change without notice, the conversation becomes a lock-picking exercise. Grim Fandango had the same disease in places — see the adventure game’s last great argument — and it’s notable that both games are remembered for their writing and forgiven for their locks.

What four CDs actually bought

The production side deserves a note, because it explains both the game’s strengths and its sag. The Longest Journey is pre-rendered 2D backgrounds with 3D characters composited on top — the Alone in the Dark trick, industrialised. In 1999 that was the pragmatic choice: real-time 3D environments looked like cardboard, and a painted still could carry detail no polygon budget of the era could touch.

The consequence is that every location costs artist-months and none of it is reusable. So a studio building a game this way has two options. Ship fewer rooms, or fill the rooms you have with talk. Funcom did both, and the four discs are mostly voice. That’s the trade at the heart of the game: the environments are static and gorgeous and finite, and the density lives in the dialogue channel because that’s the only channel that scales.

It also explains the middle. Chapters where April is stuck in one district for an hour aren’t a writing failure so much as an art-budget shape showing through the fiction. Recognising that doesn’t make the pacing better, though it does make the game legible. You’re looking at the last generation of adventures that could afford to be beautiful, spending its money exactly where the technology let it.

The real ancestor

The obvious lineage is the SCUMM doctrine: verb-driven, death-free, comedy-forward. The Longest Journey takes the no-death, no-dead-end promise and keeps it, which by 1999 was table stakes.

The more honest ancestor is the text era. This is a game whose primary verb is ask about, whose primary reward is a paragraph you didn’t have before, and whose real inventory is things April knows. Strip the pre-rendered backgrounds and you have an Infocom game with a very good narrator — which is precisely the inheritance the text parser and what we lost with it is about. Funcom’s contribution was recognising that the parser was the wrong interface for that pleasure and that a dialogue tree was the right one.

You can draw the line forward too. Disco Elysium is the same machine with better engineering: a protagonist who narrates, a world that yields to conversation, an inventory of facts. Kentucky Route Zero takes the two-worlds seam and makes it a highway. The genre didn’t die in 1999. It shed the verb coin and kept the talking.

Where to play it

It’s on PC via the usual storefronts and it runs fine on modern hardware, which is more than several of its contemporaries manage. The pre-rendered backgrounds have aged better than any polygonal work from the same year — fixed-camera 2D art doesn’t rot the way a 1999 character model does. The character models themselves are rough. You’ll adjust inside twenty minutes.

Go in expecting a novel that occasionally asks you to solve something. Turn on subtitles, click on everything, and accept that the duck is coming. If you want the whole arc, Dreamfall Chapters completes it, though it’s a different kind of game and a longer argument than this piece has room for.

Spoilers below

The plot resolves around April learning she is one of two Shifters and that the Balance between Stark and Arcadia is maintained by a Guardian at the Tower — a post that has been vacant, which is why both worlds are going wrong. The Draic Kin, the dragons, have been managing the gap. Gordon Halloway is the other Shifter, split into two halves, and the endgame is about reassembling him so he can take the Guardian’s chair.

The ending is the game’s biggest structural gamble and I think it half-works. April spends the whole story being told she is the chosen one, and the actual resolution is that her job was to deliver someone else to the throne. She hands off the destiny. Then the final scene puts her back in Arcadia, still displaced, with the sequel hook running.

The craft reading: that’s a genuinely brave beat for a 1999 adventure, because the entire genre had trained players to expect that the protagonist’s arc ends at the top of the tower. Tørnquist writes April as a person whose significance is instrumental, and then makes you feel the cost of it. The reason it only half-lands is that the last chapter starves — the exposition arrives in long speeches from characters you’ve barely met, and the game explains its cosmology at exactly the moment you want it to trust you. The mechanical version of the same problem as the duck: a grammar change at the worst time.

What holds up is April herself. Twenty-plus years on, she’s still one of the few adventure protagonists written as a young woman with a job, a landlord and a temper rather than a quipping cursor. That’s the thing the obituaries missed. The genre wasn’t out of puzzles. It was out of people worth spending fifteen hours with, and Funcom found one.