The popularity of cultural objects is often distributed as many unpopular “flops” alongside a few “hits.” Hits can be several orders of magnitude more popular than typical objects but are difficult to predict ex-ante. Most explanations focus on cumulative advantage (CA): rich-get-richer processes wherein the success of cultural objects breeds future success, creating high inequality in popularity and decoupling popularity from the properties of objects. I present an additional formal model, multiplicative returns (MR), which assumes that cultural objects are judged as interdependent ensembles of their properties. Like CA, the MR model produces “hit” objects whose popularity is difficult to predict before the fact. I show that the MR model generates popularity distributions consistent with that of US baby girls’ names. I conclude that highly unequal popularity distributions of cultural objects are over-determined: they can arise from MR, CA, or both.

文化对象的受欢迎程度通常与一些“热门”一起分布为许多不受欢迎的“失败”。命中可能比典型对象受欢迎几个数量级，但很难事前预测。大多数解释都集中在累积优势 (CA) 上：从富人到富人的过程，在这个过程中，文化物品的成功孕育了未来的成功，造成了人气方面的高度不平等，并将人气与物品的属性脱钩。我提出了一个额外的正式模型，乘法回报 (MR)，它假设文化对象被判断为其属性的相互依存的整体。与 CA 一样，MR 模型产生的“热门”对象的流行度在事前难以预测。我展示了 MR 模型生成的流行度分布与美国女婴名字的流行度分布一致。