Minimax featural computation in mathematical biolinguistics

我的研究计划源于“弱人工智能/强人工智能”的争论，这在Max Tegmark的工作中很明显，但最终源于Alan Turing的工作。我的暑期项目在音韵学领域支持了这个问题，紧跟Hale & Reiss（2008）及其对最优性理论的批评。我希望将其发展到语法领域，在这个领域，“无实质”的想法很难避免，特别是在将大量语义内容整合到应该是语法特征的制图方法中。那么，在图灵程序中提出的语言理论问题是，评估当前语言理论在多大程度上实现了强生成性的目标，并改进了这样的形式主义。正如我的提案细节，我认为这种方法与特性的精确性质以及它们如何组合有很大关系。在这个意义上，我与Cedric Boeckx（2014）有所不同，但我相信进一步的调查将揭示重要的重叠。在我看来，这必然会发展成一种强有力的跨学科方法，与生物语言学计划相一致。这些领域包括数学（参见宋2019年关于范畴论的论文）、计算机科学（复杂性理论）、认知科学（Poeppel; Gallistel & King）以及生物和进化问题（Boeckx; Balari & Lorenzo）。雄心勃勃的目标是建立一个理论，从根本上批判性地交织这些理论和学科的基本方面。

My research proposal stems from the 'weak AI/strong AI' debate evident in Max Tegmark's work but ultimately stemmingfrom the work of Alan Turing. My summer project countenanced this problem in the domain of phonology, closelyfollowing Hale & Reiss (2008) and their criticism of Optimality Theory. I wish to develop this into the domain of syntax,where the idea of being 'substance-free' is somewhat harder to avoid, especially in cartographic approaches whichincorporate a lot of semantic content into what should be syntactic features. The question posed in the Turing programfor linguistic theory is, then, to evaluate the extent to which current linguistic theories achieve the goal of stronggenerativity, and to improve formalisms as such. As my proposal details, I see this approach as having a lot to do withthe precise nature of features and how they combine. In this sense I diverge from Cedric Boeckx (2014), but I believethat further investigation will reveal important overlap. On my view, this necessarily develops into a stronglyinterdisciplinary approach, in line with the biolinguistic programme. Such areas include mathematics (cf. Song's 2019dissertation on category theory), computer science (complexity theory), cognitive science (Poeppel; Gallistel & King) andbiological and evolutionary concerns (Boeckx; Balari & Lorenzo). The ambitious goal is to build a theory from the groundup which critically intertwines the essential aspects of these theories and disciplines.