第一次真正“靠近”中国壁画，是在山西晋中的晋之源壁画艺术博物馆。
说“靠近”，并不是因为我第一次看见壁画。事实上，多年来，我去过敦煌、新疆克孜尔、山西大同，也在不同地方看过不少寺观壁画。真正的问题是：过去虽然看见了，却未必真的“看清”。
洞窟太暗，距离太远，时间太短，许多壁画又高悬于穹顶或梁壁之上。人站在现场，往往只能仰头匆匆一瞥。知道那里有飞天，有菩萨，有山水，有人物，却难以真正进入那些线条、色彩与细节之中。
晋之源壁画艺术博物馆展示的，并不是原作壁画，而是经过高清扫描后的数字输出图像。它以“中国壁画史”为脉络，将不同朝代、不同地区的经典壁画重新组织成一个可观看、可阅读、可比较的视觉空间。
有趣的是，正因为它不是原作，我反而第一次真正看见了许多过去没看见的东西。
我开始看见人物衣纹的转折，看见矿物颜料留下的层次，看见墙面裂痕之间仍未消失的笔触，看见飞天的眼神，看见永乐宫神仙队伍里不同人物的表情差异。我甚至第一次意识到，中国壁画原来有如此强烈的“运动感”。
过去在寺观或洞窟里，人们往往被一种“朝圣式”的观看方式所限制。人必须顺着空间移动，顺着光线移动，也顺着时间移动。壁画属于宗教空间的一部分，而不是单独存在的艺术对象。
但进入博物馆之后，壁画的观看逻辑改变了。它们从宗教现场进入展览空间，从建筑的一部分变成艺术史的一部分。
这让我不断思考一个问题：我们今天看到的，究竟还是“壁画”吗？或者说，它已经是一种经过技术输出转译后的“新文本”？
过去我谈的文图学，讨论的是各种形式的文本生产与诠释，而到了AIGC时代，我们开始面对另一种新的情况：任何图像，都可能经过扫描、计算、数据化与再生成。
于是，图像不再只是图像，它同时也是数据。
晋之源壁画艺术博物馆给我的震撼，恰恰正在这里。
我发现，数字技术并不只是“复制”壁画，而是在重新组织人与壁画之间的关系。过去，壁画的重点在“保存”；今天，壁画开始进入“传播”。过去，壁画强调原址；今天，壁画开始脱离原址，进入新的媒介环境。过去，观看是一种空间经验；今天，观看同时也是一种数据经验。
这些，其实都与AIGC时代的人类视觉经验非常接近。
现在许多人第一次看见艺术作品，并不是在美术馆，而是在手机屏幕上。甚至有些年轻人，是先看见AI生成的“敦煌风”“宋画风”，才开始对传统艺术产生兴趣。
有人因此担忧：数字化会不会让人越来越不在乎原作？但我在晋之源壁画艺术博物馆里，反而产生相反的感受。
正因为数字扫描让我看清了细节，我才更想去看原作。原作与数字图像之间，并不是取代关系。数字图像提供的是“可阅读性”，原作提供的则是“存在感”。高清扫描能够让人看见裂痕，却无法完全传达墙壁的厚度；能够让人看见线条，却无法替代人在空间中的身体感；能够让人看清颜色，却无法复制时间真正沉积于墙面上的气息。
真正的壁画，仍然带着一种时间重量。但数字技术却让更多人终于有机会“靠近”它。长期以来，中国壁画始终处于一种“看得见却不容易靠近”的状态。与卷轴绘画相比，壁画太巨大，也太遥远。它们存在于山中、洞窟中、寺庙中，甚至存在于普通人难以抵达的地方。某种意义上，中国壁画一直是一种“被遮蔽的视觉传统”。
然而今天，高清扫描、数字输出、人工智能图像处理技术，却开始改变这种状况。它们让壁画重新进入大众视野，也重新进入当代人的视觉经验。晋之源壁画艺术博物馆，正好提供了一个非常具体的现场。
在那里，我忽然意识到：我们终于可以靠近壁画了。
但更重要的是——当我们终于能够靠近壁画时，我们也正在重新学习，如何观看。
2026年7月4日，新加坡《联合早报》“上善若水”专栏
Finally, We Can Get Close to the Murals
I Lo-fen
The first time I truly came “close” to Chinese murals was at the
Jinzhiyuan Mural Art Museum in Jinzhong, Shanxi.
By “close,” I do not mean that it was the first time I had ever seen
murals. In fact, over the years, I have been to Dunhuang, Kizil in Xinjiang,
and Datong in Shanxi, and I have also seen many temple and monastery murals in
different places. The real issue is this: although I had seen them before, I
had not necessarily seen them clearly.
The caves were too dark, the distance too great, and the time too short.
Many murals were also high above, on domes or beams and walls. Standing on
site, one often could only look up and cast a hurried glance. One knew that
there were flying apsaras, bodhisattvas, landscapes, and figures there, but it
was difficult to truly enter into those lines, colours, and details.
What the Jinzhiyuan Mural Art Museum displays are not original murals,
but digitally printed images produced after high-definition scanning. Taking
“the history of Chinese murals” as its thread, it reorganizes classic murals
from different dynasties and different regions into a visual space that can be
viewed, read, and compared.
Interestingly, precisely because they are not the originals, I for the
first time truly saw many things that I had not seen before.
I began to see the turns of the figures’ garment folds, the layers left
by mineral pigments, the brushstrokes that had not yet disappeared among the
cracks on the wall surface, the eyes of the flying apsaras, and the differences
in expression among the various figures in the procession of immortals at
Yongle Palace. I even realized for the first time that Chinese murals possess
such a powerful sense of movement.
In the past, in temples, monasteries, or caves, people were often
restricted by a kind of “pilgrimage-style” mode of viewing. One had to move
according to the space, according to the light, and also according to time.
Murals belonged to a religious space; they were not independent art objects.
After entering the museum, however, the logic of viewing the murals
changed. They moved from religious sites into exhibition spaces, and from parts
of architecture into parts of art history.
This made me keep thinking about one question: what we see today—are
they still “murals”? Or have they already become a kind of “new text”
translated through technological output?
In the past, the Text and Image Studies I discussed concerned the
production and interpretation of texts in various forms. But in the age of
AIGC, we have begun to face another new situation: any image may undergo
scanning, computation, datafication, and regeneration.
Thus, an image is no longer merely an image. It is also data.
The shock that the Jinzhiyuan Mural Art Museum gave me lies precisely
here.
I discovered that digital technology is not merely “copying” murals, but
reorganizing the relationship between people and murals. In the past, the
emphasis of murals lay in “preservation”; today, murals have begun to enter
“dissemination.” In the past, murals emphasized the original site; today, they
have begun to detach from the original site and enter new media environments.
In the past, viewing was a spatial experience; today, viewing is also a data
experience.
All of these are in fact very close to human visual experience in the
age of AIGC.
Today, many people first encounter works of art not in art museums, but
on mobile-phone screens. Some young people even first see AI-generated
“Dunhuang style” or “Song painting style” images before they begin to take an
interest in traditional art.
Some people therefore worry: will digitization make people care less and
less about the originals? But in the Jinzhiyuan Mural Art Museum, I had the
opposite feeling.
Precisely because digital scanning allowed me to see the details
clearly, I wanted even more to see the originals. The relationship between
originals and digital images is not one of replacement. What digital images
provide is “readability”; what originals provide is “presence.” High-definition
scanning can allow people to see cracks, but it cannot fully convey the
thickness of a wall; it can allow people to see lines, but it cannot replace
the bodily sensation of being in a space; it can allow people to see colours
clearly, but it cannot reproduce the atmosphere of time truly accumulated on
the wall surface.
True murals still carry a certain weight of time. Yet digital technology
has allowed more people finally to have the chance to “come close” to them. For
a long time, Chinese murals have remained in a state of being “visible yet not
easy to approach.” Compared with scroll paintings, murals are too vast and too
distant. They exist in mountains, caves, temples, and even in places difficult
for ordinary people to reach. In a certain sense, Chinese murals have always
been a “concealed visual tradition.”
Today, however, high-definition scanning, digital output, and artificial
intelligence image-processing technologies have begun to change this situation.
They allow murals to re-enter public view and to re-enter the visual experience
of contemporary people. The Jinzhiyuan Mural Art Museum provides a very
concrete site for this.
There, I suddenly realized: we can finally get close to murals.
More importantly—when we can finally get close to murals, we are also
relearning how to look.
4 July 2026, Singapore, Lianhe Zaobao, “Shangshan Ruoshui” column
