No, it is not a bug in Selenium. You cannot select a text node in Selenium using XPath. Even though such an XPath query would be valid on an XML file (for use in an XSLT transform, for instance), HTML/DOM is not equivalent to XML.
A brief history lesson: XPath was not intended to be an element location strategy for HTML documents. It was intended to be used for XML documents. While there are some similarities between HTML and XML (both are roughly hierarchical in structure), they are not identical. HTML in particular is much looser in its allowed markup formatting than the much stricter XML. Therefore, when using Selenium WebDriver’s findElement using an XPath locator, the result must resolve to an element, not any other type of (otherwise valid) XPath node type, like an attribute node or a text node.