I have not found a document explaining my problem below. In my understanding, if \pnode(1,1){A} then \pscircle[origin={A}](A){1} must be identical to
\pscircle[origin={1,1}](A){1}\pscircle[origin={A}](1,1){1}\pscircle[origin={1,1}](1,1){1}
However, I got the following result that is different from my mental model above. How does PSTricks make assumption when handling nodes and literal points?

\documentclass[border=12pt]{standalone}
\usepackage{pst-node}
\psset{linecolor=red}
\def\xxx{\hspace{5mm}}
\begin{document}
\begin{pspicture}[showgrid=top](3,3)
\pnode(1,1){A}
\pscircle[origin={A}](A){1}
\end{pspicture}\xxx
\begin{pspicture}[showgrid=top](3,3)
\pnode(1,1){A}
\pscircle[origin={1,1}](A){1}
\end{pspicture}\xxx
\begin{pspicture}[showgrid=top](3,3)
\pnode(1,1){A}
\pscircle[origin={A}](1,1){1}
\end{pspicture}\xxx
\begin{pspicture}[showgrid=top](3,3)
\pscircle[origin={1,1}](1,1){1}
\end{pspicture}
\end{document}


pstricksmanual be a credible source for the answer? – percusse Jul 30 '12 at 22:04originand literal points. – Click Me Jul 31 '12 at 3:44