他们什么都没加,只抽走了「上」
They Added Nothing. They Withdrew "Up".
Amir Bastan、Noor Stenfert Kroese、Johannes Braumann 三个人在林茨艺术大学的 Creative Robotics 做了一件叫 MycoGravity 的装置。一台 KUKA Iontec K30R2100 六轴工业机械臂,末端挂着一个自制的生物反应器,里面长着粉红平菇(Pleurotus djamor)的菌丝。机械臂不停地换姿态——他们管这套走法叫量化随机游走,取自 NASA 的随机定位机:在三维空间里挑一批点,一边多轴旋转一边在点之间过渡。反应器内部埋着针状电极记录菌丝的生物电,外层是温度、湿度、CO₂、光照和惯性测量单元,数据实时送到一台本地服务器,服务器再回过头去调机械臂的动作和舱内环境。
要点在于他们没有消除重力。重力还在,1g,一点没少。被拿走的是方向。这个原理叫矢量平均:重力矢量被持续重新定向,从不在任何一个方位停留到足够长,长到菌能对齐它。他们还特意说明,这个效果不靠转速——慢速的多轴旋转就足以打断真菌的向重力性。
先看这里被凿的是什么。向重力性是生物学里最彻底的已构之一。菌褶朝下,胚芽朝上,根朝下——现象被观察、被命名、写进每一本教科书,凿构循环在一百多年前就跑完了。它现在是一句话,一个术语,一张示意图。这上面已经没有可凿的地方。而 Bastan 他们做的不是往体系里再加一个变量。他们抽掉了参照本身。这一次的凿子是负的:它不落在材料上,落在材料赖以成形的那个前提上——那个从来不被当成前提、因为它太稳定所以从不被写进方程的东西。
于是长出来的东西,既不是设计也不是随机。它是一套完好无损的感知系统在输入无解时留下的残迹。菌丝的向重力性没有坏,电极还在记录,它一直在测,一直在朝一个持续移动的下方对齐——所以形态是那次持续对齐的沉积物。这个东西没有名字。你可以把它塞进"生物艺术",但那是 1990 年代的构:转基因、组织培养、Kac 的兔子,一整套关于操纵生命的伦理剧场。MycoGravity 什么都没操纵,它只是撤走了一个坐标系。你也可以塞进"机器人艺术",但机械臂在这里不是表演者,是一台负责让某件事不发生的机器。两个格子都是现成的,都不合身。
他们的 FAQ 里有一条诚实到有点动人的问题:「我现在为什么看不到蘑菇?」——有时候舱里只有菌丝在活动,没有子实体。作品是空的,装置照常运行,照常产出数据。你看到什么,取决于你什么时候来。这个空不是故障,也不是禅意,它就是余项本身:作品的产出节奏不归展期管。还有:蘑菇不被采、不被吃,展完晒干保存,作为"它们如何回应陌生力量的物理记录"。一个干标本,同时是雕塑、是数据、是一次没能完成的定向。
现在看比以后看要紧,理由写在他们自己的说明里:真菌究竟怎么感知重力,"在很大程度上仍然未知"。这句话是余项还活着的证明——问题还没有答案,所以还没有构。而收编已经开始了。2025 年在 Ars Electronica 展出,2026 年 8 月 CreativeApplications 报道,TU Delft 写过,会议论文也有了。太空农业、astromycology、"真菌生物计算"这几个格子都已经在招手,而且都快满了。一旦 MycoGravity 变成"微重力作物研究的一个艺术示范",它就是构了:正确,好看,不再生长。现在去看,看到的还是那个更奇怪的东西——三个人造了一台重达数百公斤的工业机器,唯一的功能是让一种生物永远无法确定哪边是上,然后耐心地记录它继续尝试。
amirbastan.com · MycoGravity ↗ creativeapplications.net ↗Amir Bastan, Noor Stenfert Kroese and Johannes Braumann have built an installation called MycoGravity at Creative Robotics, University of Arts Linz. A KUKA Iontec K30R2100 six-axis industrial arm carries a custom bioreactor on its end effector, and inside the bioreactor pink oyster mycelium (Pleurotus djamor) is growing. The arm never settles. They call its path a quantised random walk, borrowed from NASA's random positioning machines: pick points in three-dimensional space, rotate across multiple axes while transitioning between them, repeat. Subdermal needle electrodes read the mycelium's bioelectric activity; the outer shell carries temperature, humidity, CO₂, light and an IMU. All of it streams to a local server, which feeds back into both the robot's motion and the conditions in the chamber.
The point is that gravity is not removed. It is still there, still 1g, undiminished. What is removed is direction. The principle is called vector averaging: the gravitational vector is continuously reoriented so that it never holds one bearing long enough for the fungus to align to it. They are careful to add that the effect does not depend on speed — slow multi-axial rotation is enough to disrupt gravitropic sensing.
Look at what is being chiselled here. Gravitropism is one of the most completely finished constructs in biology. Gills face down, shoots go up, roots go down: observed, named, written into every textbook, the chisel-construct cycle closed more than a century ago. It is now a sentence, a term, a diagram. There is nothing left in it to cut. And what Bastan and his collaborators do is not add another variable to the system. They withdraw the reference itself. The chisel this time is negative: it does not land on the material, it lands on the premise the material forms against — the thing that is never treated as a premise, never written into the equation, precisely because it is too stable to notice.
So what grows is neither designed nor random. It is the residue of an intact sensing system running on an input that has no solution. The mycelium's gravitropism is not broken. The electrodes keep recording. It keeps measuring, keeps aligning to a down that keeps moving — and the resulting form is the sediment of that continuous aligning. This thing has no name. You could file it under bio-art, but bio-art is a 1990s construct: transgenics, tissue culture, Kac's rabbit, an entire ethical theatre about manipulating life. MycoGravity manipulates nothing. It withdraws a coordinate system. You could file it under robotic art, but the arm is not a performer here; it is a machine whose job is to make something not happen. Both slots already exist. Neither one fits.
Their FAQ contains a question that is honest to the point of being moving: "Why don't I see mushrooms right now?" Sometimes there is only mycelium in the chamber and no fruiting bodies. The work is empty; the installation runs anyway, produces data anyway. What you see depends on when you visit. That emptiness is not a fault and it is not a piece of Zen — it is the remainder itself: the work's output is not on the exhibition's schedule. And afterwards the mushrooms are neither harvested nor eaten. They are dried and kept, as a physical record of how they responded to unfamiliar forces. A dried specimen that is simultaneously sculpture, data, and an orientation that never resolved.
Seeing it now matters more than seeing it later, and the reason is in their own text: how fungi sense gravity is "still largely unknown." That sentence is the proof that the remainder is alive — the question has no answer yet, so there is no construct yet. And the absorption has already started. Ars Electronica showed it in 2025; CreativeApplications covered it in August 2026; TU Delft has written about it; there is a conference paper. Space agriculture, astromycology, fungal biocomputing — those slots are all waving, and they are all filling fast. The moment MycoGravity becomes "an artistic demonstration for microgravity crop research," it is construct: correct, handsome, no longer growing. Go now and you still see the stranger thing — three people built a several-hundred-kilogram industrial machine whose only function is to make an organism permanently unable to determine which way is up, and then patiently recorded it going on trying.
amirbastan.com · MycoGravity ↗ creativeapplications.net ↗