哈贝马斯之后 After Habermas

 

那天早上，我在手机上刷到消息：哈贝马斯（Jürgen Habermas, 1929–2026）走了。

我呆了一下。不是因为意外——他96岁，已是长寿，而是有一种说不清楚的感觉，像是什么东西跟着他一起，悄悄离开了。

我想到的，不只是世界少了一位大哲学家，而是：那个相信人类还能通过讨论、说理、思辨来支持共同生活的时代，好像也在慢慢收尾了。

哈贝马斯提出的重要概念之一，是"公共领域"(public sphere)。他认为：在家庭、市场、国家机器之外，社会该还有一个中间地带，让人们可以围绕共同关心的事情，好好地交换意见、形成舆论。在公共领域，不是比谁的声音大、权力大，而是人们彼此理性沟通，愿意把话讲清楚，也愿意听别人讲清楚。

这听起来简单朴素，做起来却很难,而且越来越难。

我有时候浏览手机，刷着刷着，会有一种空虚的疲倦感——信息源源不绝；影音眼花缭乱。人人都在急切地表达，却不一定有人在倾听。人人都在转发，却不一定有人会仔细看。

哈贝马斯担心的“公共领域的衰落”，以一种令人始料未及的方式发生着。公共讨论没有消失，只是被算法和流量悄悄改变了状态。

然后，AI来了。

现在不只是人在说话，机器也在生成内容。一段文字、一张图、一段声音，可以在几秒内被产出、改写、包装、扩散。我们的公共空间，已经不再只是"谁在讲话"的问题，而是"谁在生成""谁在分发""谁在决定什么会被看见"的问题。

这正是我这几年一直在研究和教学AIGC文图学（Text and Image Studies on AIGC）的原因。

每次提到AIGC（人工智能生成文本），很多人第一反应是便利的工具。但我更在意的，是另一个方向：当文本和图像越来越容易被机器生成，人要怎么理解这些内容？怎么判断它们？怎么看出它们背后藏着什么样的结构、情绪、立场？

这其实和哈贝马斯在意的事情十分接近。

他关心的是：人如何通过语言沟通，建立共同的理解。我关心的是：在一个图像泛滥、内容爆炸、AI无处不在的时代，人还有没有可能维持共同理解的基础。他那个时代，公共讨论的主要战场是报纸、辩论、演讲；今天，这个战场已经延伸到了互联网、充斥短视频、表情包、AI生成的各种多模态文本。

这也让我重新省思，我为什么要创立文图学会。

创立文图学会，是希望通过多模态文本与跨媒介表达，促进公共理解与文化对话。换句话说，文图学会不只是学术团体，也是一个把学术研究、文化传播、媒介表达与社会沟通连接起来的平台。文图学会推动的，不只是扩充知识，分享艺文，还要理解知识如何被生产、被继承、被诠释。让复杂的论述可以被更多人知晓，让不同背景的人能够参与，让学术不只停留在专业圈里，而能够转化为公共语言，普及于社会。

如果借用哈贝马斯的眼光来看，文图学会所做的事情，某种程度上正是在今天新的媒介环境里，努力为营造公共领域的空间。这个空间并非传统意义上“坐下来辩论”的场所，而是经由文字、图像、影音、课程、讲座、出版，以及现在的 AIGC 实践，让我们重新学习“万物皆文本，文本可生成”，重新认识真实与人工智能生成的世界。

哈贝马斯去世，对我来说不只是一个新闻，而更像一种提醒：哈贝马斯之后，真正重要的，不是怀念过去那个哲学思考的光华，而是在这个人人都在表达，连机器也在表达的时代，我们还愿不愿意，也还能不能理性交流——真正听见彼此。

 

2026年6月6日，新加坡《联合早报》“上善若水”专栏

 

After Habermas

I Lo-fen

That morning, I saw the news on my phone: Jürgen Habermas (1929–2026) had passed away.

I paused for a moment. Not because it was unexpected—he was ninety-six and had lived a long life—but because of an indefinable feeling, as though something had quietly departed along with him.

What came to mind was not merely that the world had lost a great philosopher. It was also this: the era that believed human beings could still sustain a shared life through discussion, reasoned argument and reflection seemed to be drawing slowly to a close.

One of Habermas’s most important concepts was the “public sphere.” He believed that beyond the family, the market and the machinery of the state, society should have an intermediate space where people could exchange views on matters of common concern and form public opinion. In the public sphere, what mattered was not who had the loudest voice or the greatest power, but whether people could communicate rationally, explain themselves clearly and be willing to listen as others did the same.

This sounds simple and straightforward. In practice, however, it is difficult—and becoming increasingly so.

Sometimes, as I scroll through my phone, I feel a hollow weariness. Information flows without end; images and videos flash before the eyes. Everyone is eager to express themselves, but not necessarily to listen. Everyone is forwarding content, but not necessarily reading or watching it carefully.

The “decline of the public sphere” that Habermas feared is taking place in a way few could have anticipated. Public discussion has not disappeared; rather, its condition has been quietly transformed by algorithms and traffic.

Then AI arrived.

Now it is no longer only human beings who speak. Machines generate content as well. A piece of writing, an image or a voice recording can be produced, rewritten, packaged and disseminated within seconds. Our public space is no longer concerned merely with “who is speaking,” but also with “who is generating,” “who is distributing” and “who decides what will be seen.”

This is precisely why I have spent the past several years researching and teaching Text and Image Studies on AIGC.

Whenever AIGC—AI-generated content—is mentioned, many people first think of it as a convenient tool. What concerns me more, however, is another question: when texts and images can be generated by machines with increasing ease, how should people understand such content? How should they judge it? How can they discern the structures, emotions and positions concealed behind it?

This is, in fact, very close to what concerned Habermas.

He asked how human beings could establish mutual understanding through linguistic communication. I ask whether, in an age flooded with images, overwhelmed by content and permeated by AI, it is still possible to preserve a basis for shared understanding. In his time, the main arenas of public discussion were newspapers, debates and speeches. Today, that arena has expanded to the internet, crowded with short videos, memes and all manner of multimodal texts generated by AI.

This has also prompted me to reflect anew on why I founded the Text and Image Studies Society.

The Society was established in the hope of promoting public understanding and cultural dialogue through multimodal texts and cross-media expression. In other words, it is not merely an academic organisation, but a platform that connects scholarly research, cultural communication, media expression and social dialogue. What it seeks to promote is not only the expansion of knowledge and the sharing of arts and culture, but also an understanding of how knowledge is produced, inherited and interpreted. It aims to make complex arguments accessible to more people, to enable those from different backgrounds to participate, and to ensure that scholarship does not remain confined to professional circles but can be transformed into a public language and disseminated throughout society.

Seen through Habermas’s perspective, the work of the Text and Image Studies Society may, in a certain sense, be understood as an effort to create a space for the public sphere within today’s new media environment. This space is not a conventional setting in which people simply “sit down and debate.” Rather, through writing, images, audiovisual media, courses, lectures, publications and now the practice of AIGC, it allows us to relearn that “everything is text, and texts can be generated,” and to understand anew both the real world and the world generated by artificial intelligence.

For me, Habermas’s death was not merely a piece of news. It was more like a reminder. After Habermas, what truly matters is not nostalgia for the brilliance of a past age of philosophical thought. What matters is whether, in an era when everyone is expressing themselves and even machines have begun to speak, we are still willing—and still able—to communicate rationally and truly hear one another.

Originally published in the “Shang Shan Ruo Shui” column of Singapore’s Lianhe Zaobao, 6 June 2026.

“晓寒”还是“晚归” “Morning Chill” or “Returning at Dusk”?

 

新加坡国家美术馆正在展出“何香凝：画就丹青凭寄意”展，主视觉是一头描绘细腻的狮子。以前我看过的何香凝（1878-1972）画作主要是花鸟和山水，知道她曾经到过新加坡开画展，心想也许会有一些和新加坡有关的故事。

果然，看到她1929年在丘菽园收藏的《林则徐墨宝》上题字“努力”，那虫洞斑驳的纸面，真令人心疼。同样在 1929 年，何香凝访问新加坡之前的二月间，陈树人（1884-1948）在上海画的一幅《牡丹双蝶图》吸引了我和学生的目光。

花枝斜出，两只蝴蝶翩然，笔致清秀，设色淡雅，典型的陈树人艺术风格。画面右上方大片留白，正有题诗一首。整幅画匀称，通透，给人轻盈之感，一眼看去，只觉得是一个美丽的春天。

然而仔细读展签上录的题诗，发现有些不对劲："春残露华重，晚归香满枝。可怜刘碧玉，娇小嫁人时。"再对照画上陈树人的行书，首二句写的分明是"春残露初重，晓寒香满枝"，不是"晚归"。落款那天是花朝节，百花生日，清晨露重香满。"晓寒"，是花朝清晨那一刻薄寒未退、满枝盈香的气息；改成"晚归"，时分从拂晓移到黄昏，花朝的意涵散了，诗的情境也变了。

说来有趣，我在画上辨认那个"初"字时，颇费了些工夫。陈树人把这个字写成上下结构，上“衣”下“刀”（看起来像“力”字）。不似常见的左右结构，我盯着看了许久，第二天才恍然——原来是"初"啊。一个字，让人悬想了整整一天。

题诗的来历，要从居廉说起。居廉（1828-1904）是陈树人与高剑父共同的老师，岭南画派的重要奠基者。1886年，居廉在《杂花册》中题写了同样的诗，并在末尾注明"匏谷句"——匏谷可能是晚明万历进士，广东海阳诗人吴殿邦，他的诗集即名《匏谷诗集》。

后来高剑父在1913年《芍药图》题款里沿用了"可怜刘碧玉，娇小嫁人时"；师弟陈树人1929年又将全诗题于《牡丹双蝶图》，师徒三人，同一首诗，辗转题在不同的春花画作上。这不是巧合，而是因袭，一种在岭南画派内部流通的程式，一套关于春花、关于美人、关于花朝时节的固定表达，从明代流向民国画家，像一个不成文的惯例，绵延传递。

诗里的"刘碧玉"出自乐府《碧玉歌》，本无姓氏，只是汝南王的小家爱妾，以"碧玉"为名，形容她的温润可人。至北周庾信，才写下"定知刘碧玉，偷嫁汝南王"——因她从属刘姓王侯，遂冠以"刘"姓，又以"偷"字点出小家女悄然归属的命运。牡丹在花朝节清晨无声绽放，露重香满，与小家碧玉娇小嫁人时的盛美与脆弱，是同一种美丽的瞬间。诗人以花喻人，以人喻花，两者合而为一。

然而回头看这幅画，画面上有两只蝴蝶，诗里却没有蝴蝶的影子。图像与文字并不对应，各自讲述自己的事。这正是文图学（Text and Image Studies）感兴趣的地方：在中国书画传统里，诗书画合一的模式早在宋代已经成熟，画家在构图时便为题诗预留空间，文字是画面整体的一部分，而非事后填补。但这并不意味着诗与画必须在内容上彼此呼应。题诗带来的是另一个意义维度，让观者在蝶与花之外，忽然读到一个小家女出嫁的清晨，似乎两个世界并置，不解释，也不需要解释。

只是，当诗的文字出了错，这个并置便悄悄走了形。"晓寒"变成"晚归"，清晨变成黄昏，一首在岭南画派流传的诗，在一张展签里，静静地讹传下去。而看画的人，大约只记得那两只翩飞的蝴蝶，以及满纸春光。

 

2026年5月23日，新加坡《联合早报》“上善若水”专栏

 

“Morning Chill” or “Returning at Dusk”?

I Lo-fen

The National Gallery Singapore is currently presenting He Xiangning: Painting with Resolve, whose key visual features a meticulously rendered lion. The works by He Xiangning (1878–1972) that I had previously seen were mostly flower-and-bird paintings and landscapes. Knowing that she had once held an exhibition in Singapore, I wondered whether the show might contain stories connected with the city.

Sure enough, I came across the word “Strive,” inscribed by He in 1929 on a piece of Lin Zexu’s calligraphy from the collection of Khoo Seok Wan. The paper, mottled and riddled with wormholes, was truly painful to behold. Also in 1929, in February, shortly before He Xiangning visited Singapore, Chen Shuren (1884–1948) painted a work in Shanghai titled Peonies and Two Butterflies. It immediately caught the attention of both my students and me.

A flowering branch extends diagonally across the composition as two butterflies flutter beside it. The refined brushwork and restrained, elegant colours are typical of Chen Shuren’s artistic style. A large expanse of blank space occupies the upper-right corner, where a poem has been inscribed. The composition is balanced, open and airy, conveying an overall impression of lightness. At first glance, it seems simply to depict a beautiful spring day.

Yet when I read the poem transcribed on the exhibition label, something seemed amiss:

As spring wanes, the dew lies heavy;
returning at dusk, fragrance fills the branches.
How endearing is Liu Biyu,
petite and lovely at the time of marriage.

When I compared this transcription with Chen Shuren’s running-script inscription on the painting itself, however, the first two lines clearly read:

As spring wanes, the first dew lies heavy;
in the morning chill, fragrance fills the branches.

They do not say “returning at dusk.” The painting was dated to the Flower Festival, traditionally regarded as the birthday of all flowers—a morning of heavy dew and branches laden with fragrance. “Morning chill” evokes precisely that moment at dawn when the lingering cold has not yet dispersed and the branches are filled with scent. Replacing it with “returning at dusk” shifts the scene from daybreak to evening, dissolving the significance of the Flower Festival and transforming the poem’s entire atmosphere.

Interestingly, I had considerable difficulty identifying the character chu—“first”—in the inscription. Chen Shuren wrote it in an unusual vertically stacked form, with “clothing” above and “knife” below, the latter resembling the character for “strength.” Unlike the familiar left-right structure, this form held my gaze for a long time. It was only the following day that I suddenly realised: it was chu. A single character left me pondering for an entire day.

The origins of the poem lead us back to Ju Lian (1828–1904), the shared teacher of Chen Shuren and Gao Jianfu and an important founding figure of the Lingnan School of painting. In 1886, Ju inscribed the same poem in his Album of Miscellaneous Flowers, adding at the end that the lines were by “Paogu.” Paogu may have been Wu Dianbang, a poet from Haiyang, Guangdong, who passed the metropolitan civil-service examination during the Wanli reign of the late Ming dynasty. His collected poems were titled The Paogu Poetry Collection.

Later, Gao Jianfu quoted the lines “How endearing is Liu Biyu, petite and lovely at the time of marriage” in the inscription on his 1913 Peony. In 1929, his junior fellow-student Chen Shuren inscribed the entire poem on Peonies and Two Butterflies. The same poem thus passed among three members of successive generations, appearing on different paintings of spring flowers. This was no coincidence, but a form of artistic inheritance—a convention circulating within the Lingnan School, a fixed vocabulary for spring blossoms, beautiful women and the Flower Festival. Flowing from the Ming dynasty into the Republican period, it was transmitted like an unwritten custom.

The “Liu Biyu” mentioned in the poem derives from the Music Bureau ballad Song of Biyu. Originally, the woman had no surname. She was merely a beloved concubine of the Prince of Runan from a modest family, and “Biyu,” or “green jade,” was used to suggest her gentle and pleasing beauty. It was not until the Northern Zhou poet Yu Xin wrote, “Surely Liu Biyu secretly married the Prince of Runan,” that she acquired the surname Liu. Because she belonged to a prince of the Liu clan, she was assigned his surname, while the word “secretly” hinted at the quiet destiny of a young woman of humble birth entering another household.

The peony blooms silently on the morning of the Flower Festival, heavy with dew and filling the branches with fragrance. Its beauty is as radiant and fragile as that of the petite Biyu at the moment of marriage. The poet uses the flower to evoke the woman and the woman to evoke the flower, until the two become one.

Yet when we return to the painting, we see two butterflies, although the poem contains no trace of them. Image and text do not correspond directly; each tells its own story. This is precisely what interests Text and Image Studies. In the Chinese painting-and-calligraphy tradition, the integration of poetry, calligraphy and painting had already matured by the Song dynasty. Painters often reserved space for inscriptions while planning the composition, making words an integral part of the image rather than a later addition.

This does not mean, however, that poem and painting must correspond in subject matter. The inscription introduces another dimension of meaning. Beyond the butterflies and flowers, the viewer suddenly encounters the dawn on which a young woman of humble birth is married. Two worlds seem to stand side by side, without explanation and without any need to be explained.

But when the words of the poem are misread, that juxtaposition is subtly distorted. “Morning chill” becomes “returning at dusk”; dawn becomes evening. A poem passed down within the Lingnan School thus quietly continues its altered transmission on a museum label. Most viewers, perhaps, will remember only the two fluttering butterflies and a page filled with the radiance of spring.

Originally published in the “Shang Shan Ruo Shui” column of Singapore’s Lianhe Zaobao, 23 May 2026.