Unconventional Accuracy; Objectivity in the Digital Turn

建筑的第一次数字化转变允许从业者通过计算机辅助设计和模拟将日常模拟设计过程（如绘图和建模）转换为高度具体的数字技术。第二次数字化转变要求我们从概念上质疑这些技术的作用，正如Mario Capro所强调的，新的认知技术正在进入建筑学科，也在“训练我们以不同的方式思考……[但是]除非我们学会应对这门新科学，否则我们很可能把它搞得一团糟。尽管数字技术现在在建筑实践和学术界都很常见，但作为设计师，我们仍然很少选择质疑这些高度技术过程的基本基础。我们的数字建筑工具大多来自科学和计算背景，仍然在努力寻找独特的术语，以便在空间上进行讨论。即使在最常见的可转移建筑术语中，如“渲染”，建筑师也会无意识地简化这些数字软件的潜力，这些软件能够准确地模拟大量轻粒子撞击复杂的笛卡尔几何形状。这种简化导致审美理想与类似的手绘“渲染”术语相混淆，然后被设计师的传统期望所偏见。正如西蒙·彭尼（Simon Penny）所强调的，在计算领域，拟人化定义同样被挪用，“当任何这样的断言不再被明确地理解为隐喻时，我们已经陷入了本体论的泥潭。”“[2]由于缺乏真正适用的术语，在我们长期存在的模拟语言和相对较新的数字语言之间，我们加剧了存在于建筑环境中具象思想和物理构造现实之间的问题裂痕。这个问题导致了审美意象的优先级和制造，而不是建造的建筑，并使我们处于话语的关键点；我们必须定义使用这些工具的条款，以避免陷入更深的数字泥潭。这项研究将我们当前有问题的数字裂痕与19世纪末和20世纪初科学图像制作的客观转向相提并论。在亚瑟·沃辛顿（Arthur Worthington）的带领下，与其他科学实践者一起，这一概念转变开始质疑事实科学生产中的审美偏见做法，正如彼得·加利森（Peter Galison）和洛林·达尔斯顿（Lorraine Dalston）在谈到沃辛顿的实验时所指出的那样，“追踪和记录基本的、理想的‘自动飞溅’，曾经是一种高阶科学美德，现在却变成了一种心理缺陷，一种感知缺陷。”这种对他们方法论中潜在错误的理解导致了一种概念上的转变，从对作者预期现实的偏见表征，到通过无偏见的视角观察到的真实构建、模拟和记录的现实。这种方法在当时被认为是非常规的，它剥夺了对理想完美形象的追求，这是令人钦佩的，但也是美德。这项以设计为主导的研究旨在开发实验性和非常规的实践形式，以及与我们建筑对话中未定义的数字工具相关的自我决定的准确性措施。研究主要着眼于前面提到的新形式的认知机器学习，这在我们的建筑词汇中是无与伦比的，研究将涉及有能力定义自己独特的自学空间条件理解的工具和系统。M. Carpo, 2017。第二次数字化转型。剑桥:麻省理工学院。p162。[2]彭妮，南卡罗来纳州，2019。创造意义：认知、计算、艺术和具体化。麻省理工学院出版社。p4。[3]达斯顿，L.和加里森，P., 2010。客观性。带书。p16。

The first digital turn in architecture allowed practitioners to transition everyday analogue design processes, such as drawing and modelling, to highly specific digital technologies through computer aided design and simulation. The second digital turn then asked us to question the role of these technologies conceptually and as Mario Capro highlights, new cognitive technologies are emerging into the architectural discipline that are also 'training us to think in a different way, ... [but] unless we learn to cope with this new science, we are likely to make a terrible mess of it.'[1] And even though digital technologies are now commonplace, within both architectural practice and academia, as designers we still rarely choose to question the fundamental underpinnings of these highly technical processes. Our digital architectural tools, which are mostly appropriated from scientific and computational backgrounds, still struggle to find unique terminologies by which to be discussed spatially. In even the most common transferable architectural terms, such as 'render', architects unconsciously simplify the potential of these digital softwares, which have the ability to accurately simulate vast arrays of light particles crashing against complex cartesian geometry. This simplification results in conflated aesthetic ideals that register with analogue hand drawn 'rendered' terminologies and are then biased by a conventional expectation of the designer. As Simon Penny highlights on the same appropriation of anthropomorphic definitions in computing, 'When any such assertion is no longer explicitly understood as metaphor, we are already deep in an ontological mire.'[2] As a result of this lack of truly applicable terminology, between our longstanding analogue and relatively new digital languages, we exacerbate the problematic rift that exists between representational ideas and their physically constructed reality in the built environment. This problem has led to the prioritisation and fabrication of aesthetic imagery over constructed architecture and leaves us at a critical point in our discourse; where we must define the terms of engagement with these tools to avoid descending deeper into the digital mire. This research draws parallels to our current problematic digital rift with the late 19th and early 20th century objective turn in scientific image making. Led by Arthur Worthington, alongside fellow scientific practitioners, this conceptual turn began to question aesthetically biased practices in the production of factual science, and as Peter Galison & Lorraine Dalston note with regards to Worthington's experiments, 'What had been a high-order scientific virtue - tracking and documenting the essential, ideal "auto-splash" - became a psychological fault, a defect in perception.'[3] This understanding of the underlying fault in their methodology resulted in a conceptual shift away from bias representation of an expected reality of the author, to one of truthfully constructed, simulated and documented reality that was observable through a non-biased lense. This approach, which at the time was considered unconventional, stripped away the admirable, but virtues, pursuit of the idealised perfect image. This design-led research sets out to develop experimental and unconventional forms of practice, alongside self determining measures of accuracy, in relation to digital tools that are under-defined within our architectural dialogue. Predominantly looking towards aforementioned new forms of cognitive machine learning, which are unparalleled in our architectural lexicon, the research will engage with tools and systems that have an ability to define their own unique self taught understandings of spatial conditions. [1] Carpo, M., 2017. The Second Digital Turn. Cambridge: MIT. p162.[2] Penny, S., 2019. Making Sense: Cognition, Computing, Art, and Embodiment. MIT Press. p4.[3] Daston, L. and Galison, P., 2010. Objectivity. Zone Books. p16.