The Script-based Semantic Theory of Humor (Raskin 1985) was proposed as a semantic theory (Raskin 1979), around the time than other script-like developments were discussed within the AI community. It is important to understand this in order to appropriately place the SSTH within its contemporaries. Within Artificial Intelligence back then, just like now, there was a hope that computers or other intelligent agents would be able to understand human language. It was also realized that most human conversations or scenarios take place in isolation of the overall world knowledge – at least as far a computer is concerned – and while this knowledge is readily accessible to people, computers are left in the dark about what the sentences are about. A classic example is a conversation that may happen between a hostess and a customer at a restaurant (Schank and Abelson 1977). A hostess may greet a customer, and a customermay request a table. In order for a computer to gain the same understanding of this situation that a human has, it has to have access to knowledge than when a person comes to a restaurant (which also has to be defined), (s)he will have a conversation with a host(ess), who will then seat the customer’s party at a table, if it is available, possibly placing menus that will be used as a source of information from which customers may order food, a waiter(ess) will come to take an order, send information about the order to the kitchen, ordered food will get prepared, etc. While the particulars of exactly what happens may change depending on counties, types of restaurants, and many other things, once a general sequence of events is established, it is much easier to understand what, for example, “How many?” refers to in this situation. Such scenarios were referred to as scripts (see more details on the restaurant script in Schank and Abelson 1977), and were used as an attempt to somehow structure human knowledge. In a somewhat similar fashion, Minsky proposed “frames” (Minsky 1974) as a data structure for stereotypical situations. There are quite a few differences between scripts and frames – we can go broader and add schemata – but they share the same goal: representation of knowledge of stereotypical situations, allowing individual phrases to be understood within the context of such evoked

中文翻译：

---

反对、重叠和本体——对 Elliott Oring 的回应

基于脚本的幽默语义理论（Raskin 1985）被提出作为一种语义理论（Raskin 1979），大约在人工智能社区内讨论的其他类似脚本的发展的时间。重要的是要理解这一点，以便将 SSTH 适当地置于其同时代人中。在当时的人工智能中，就像现在一样，希望计算机或其他智能代理能够理解人类语言。人们还意识到，大多数人类对话或场景都是在孤立的整体世界知识的情况下发生的——至少就计算机而言——虽然人们很容易获得这些知识，但计算机对句子是什么一无所知关于。一个典型的例子是餐厅的女主人和顾客之间可能发生的对话（Schank 和 Abelson 1977）。女主人可能会向顾客打招呼，顾客可能会要求一张桌子。为了让计算机获得与人类相同的对这种情况的理解，它必须获得知识，而不是当一个人来到餐厅（这也必须定义）时，他将进行对话与主人（ess）一起，然后将客户的聚会安排在一张桌子上，如果有的话，可能会放置将用作客户可以从中订购食物的信息来源的菜单，服务员（ess）将来到接受订单，将有关订单的信息发送到厨房，将准备订购的食物等。 虽然具体发生的事情可能会因县，餐厅类型而异，以及许多其他事情，一旦确定了事件的一般顺序，就更容易理解什么，例如，“有多少？” 指在这种情况下。此类场景被称为脚本（参见 Schank 和 Abelson 1977 中有关餐厅脚本的更多详细信息），并被用作以某种方式构建人类知识的尝试。以某种类似的方式，明斯基提出了“框架”（Minsky 1974）作为陈规定型情况的数据结构。脚本和框架之间存在相当多的差异——我们可以更广泛地添加模式——但它们具有相同的目标：表示陈规定型情况的知识，允许在此类诱发的上下文中理解单个短语 “多少？” 指在这种情况下。此类场景被称为脚本（参见 Schank 和 Abelson 1977 中有关餐厅脚本的更多详细信息），并被用作以某种方式构建人类知识的尝试。以某种类似的方式，明斯基提出了“框架”（Minsky 1974）作为陈规定型情况的数据结构。脚本和框架之间存在相当多的差异——我们可以更广泛地添加模式——但它们具有相同的目标：表示陈规定型情况的知识，允许在此类诱发的上下文中理解单个短语 “多少？” 指在这种情况下。此类场景被称为脚本（参见 Schank 和 Abelson 1977 中有关餐厅脚本的更多详细信息），并被用作以某种方式构建人类知识的尝试。以某种类似的方式，明斯基提出了“框架”（Minsky 1974）作为陈规定型情况的数据结构。脚本和框架之间存在相当多的差异——我们可以更广泛地添加模式——但它们具有相同的目标：表示陈规定型情况的知识，允许在此类诱发的上下文中理解单个短语 Minsky 提出了“框架”（Minsky 1974）作为用于陈规定型情况的数据结构。脚本和框架之间存在相当多的差异——我们可以更广泛地添加模式——但它们具有相同的目标：表示陈规定型情况的知识，允许在此类诱发的上下文中理解单个短语 Minsky 提出了“框架”（Minsky 1974）作为用于陈规定型情况的数据结构。脚本和框架之间存在相当多的差异——我们可以更广泛地添加模式——但它们具有相同的目标：表示陈规定型情况的知识，允许在此类诱发的上下文中理解单个短语