How do children construct linguistic color categories?

当孩子们学习一个新单词时，比如一个物体的名字，他们必须弄清楚哪些其他物体应该得到这个相同的标签。对于颜色词来说，这可能是一个特别具有挑战性的问题，因为同一颜色类别的不同示例可能会有很大差异，并且各个颜色标签的确切边界并不明显。例如，一旦孩子学习到蓝色可以指天空的颜色，她也必须确定蓝莓、葡萄或草是否收到相同的标签。尽管对于成年人来说，区分蓝色和紫色或绿色的界限似乎很简单，但儿童经历了一些模棱两可的事情，不同的文化实际上使用了不同的颜色类别和不同的边界。例如，一些文化指的是具有相同颜色标签的蓝色和绿色的东西。因此，要学习颜色术语，孩子们必须弄清楚蓝色的起点和终点。目前的研究通过考察英语使用者和Otji-Herero使用者的颜色词学习来调查儿童如何解决这个复杂的问题。Otji-Herero是纳米比亚农村的一个少数民族Himba人说的，它只包含5个颜色术语，涵盖了说英语的人通常使用11个术语(红色、橙色、黄色、绿色、蓝色、紫色、粉色、棕色、灰色、黑色和白色)标记的相同颜色范围。这项研究调查了接触视觉颜色的差异，以及不同大小和结构的颜色词汇，如何影响个体建立颜色词类。该项目将从三个部分调查学习者如何用语言编码颜色。首先，它将使用适合婴儿的眼球追踪方法来测试儿童的早期理解能力，以确定指导英语学习儿童(18-30个月)第一颜色词义的初始假设。第二，该项目将测试纵向类别的变化。为了做到这一点，研究人员将跟踪研究儿童，从他们产生他们的第一个颜色词的时间，直到他们获得所有11个基本英语颜色词的成人般的含义。最后，这些关于范畴变化的自然主义研究将得到实验性范畴变化研究的补充，在这些研究中，研究人员将新的颜色词引入到将其颜色空间划分为5个(Otji-Herero)或11个(如英语)词的个体中，同时操纵他们在训练过程中获得的视觉证据的类型。研究人员还将拍摄参与者的学习环境，以测试视觉体验如何影响颜色词义的变化。这些照片将被用来创建一个有价值的在线数据库，作为其他研究人员的资源。利用这些视觉图像以及现有的儿童语言输入数据库，研究人员将构建一个计算模型，说明学习者如何使用视觉和语言信息来构建颜色词义。

When children learn a new word, such as the name for an object, they must then figure out what other objects should receive this same label. This can be an especially challenging problem for color words because different examples of the same color category can vary widely and the exact boundaries of individual color labels are not obvious. For example, once a child learns that blue can refer to the color of the sky, she must also determine whether blueberries, grapes or grass receive the same label. Although the boundaries for what differentiates blue from purple or green seem straightforward to adults, children experience some ambiguity, and different cultures actually employ different color categories with different boundaries. For example, some cultures refer to blue and green things with the same color label. Thus, to learn color terms, children must figure out where blue begins and ends. The current research investigates how children solve this complex problem by examining color word learning in speakers of English and Otji-Herero. Otji-Herero is spoken by the Himba, an ethnic group in rural Namibia, and contains only 5 color terms to cover the same range of colors that English speakers typically label using 11 terms (red, orange, yellow, green, blue, purple, pink, brown, gray, black and white). The research investigates how differences in exposure to visual colors, combined with differently sized and structured color vocabularies, affect how individuals establish color word categories.The project will investigate how learners encode color with language in three parts. First, it will test children's early comprehension using infant-friendly eye-tracking methods to determine the initial assumptions that guide first color word meanings in English-learning children (18-30 months). Second, the project will test longitudinal category change. To do so, the researchers will follow children from the time they produce their first color words until they acquire adult-like meanings for all 11 English basic color terms. Finally, these naturalistic studies of category change will be complemented by studies of experimental categorical change in which researchers introduce new color words to individuals who have divided their color space into 5 (Otji-Herero) or 11 (as in English) words while manipulating the type of visual evidence they receive during training. The researchers will also photograph the learning environments of participants to test how visual experience impacts changes in color word meanings. These photographs will be used to create a valuable online database as a resource for other researchers. Using these visual images as well as existing databases on children's linguistic input, the researchers will construct a computational model of how learners use visual and linguistic information to construct color word meanings.