The name of a thing is a mere "sign" or label of its acknowledged existence, and in no case is necessarily universal in its application or meaning. While some words may be applied universally, ie: 'tree'. Its universality emerges from the further thought and not its mere name.
To imply that labels denote universals, would imply that within one word, lies the explanation. Since there is a mere impression in the mind as a result of perception, and the further mental action of annotating the object, it is a logical impossibility that a single word can be a universal. To be universally applied is not that same as being itself universal.
Besides, tree, is described in many different languages. A circularity might arise from seeing many different trees but appealing to the "form" of tree in order to categorize the trees as part of a "universal form" 'tree'. (in english)
I think Pears got a little too poetic rather than philosophical, implying that words and sentences are mere shadows and further implying that they are not separately distinguishable. I wonder how he ever wrote a sentence?
How does he understand that the proposition:
This is a dog.
Does not mean:
This is a horse.
A proposition, in logic by definition, is a statement capable of being true or false, or indeterminate. Is he saying that ALL propositions are indeterminate?
(shadowy?)
If he really felt that, then consider a conversation with him: (with simple commands)
Turn on the light.
What?
The light... turn on the light!
What's that?
Just flip that switch on the wall..
Just what, what?
Pears proposes an absurdity, in an attempt to show one.