It happens that among my personal acquaintances there is one astronomer and one psychical researcher. It further happens that these are one and the same person - Professor Archie Roy. Archie is a distinguished scholar of astronomy who retired some tears ago from his work at the University of Glasgow; he is also a long-serving member of the Scottish Society for Psychical Research. Whether he can said to 'believe in ghosts' is for Archie to say but he has written a book about numerous prima facie cases of the supposed paranormal - including ghostly appearances - under the title 'A Sense of Something Strange'. But it is not about my friend Archie that this essay is written.

Let us first define the word ghost and cognate terms. To claim 'to have seen a ghost' is ordinarily taken to mean that the witness claims to have seen things or events - including animals, people and inanimate objects- that can reasonably be assumed to have ceased to exist in the ordinary day-to-day sense of 'exist'.

Strictly within all the terms of this definition we can suppose that inanimate things can be part of ghostly appearances - ghost stories do not usually refer to naked ghosts so, if ghosts there be, clothing must be potentially as ghostly as flesh.

Strictly within the terms of this definition, but restricted only to inanimate things and events, it is clear that astronomers, as astronomers, do normally believe in ghosts. It is normal for astronomers to refer to observed things and events located evidently at a distance of, say, half a million light years from the observer. The observer is witnessing something that happened half a million years ago and, on the reasonable assumption that whatever it was is no longer there (in the case of a thing) or over and done with (in the case of an event) that observer is seeing ghosts of past things, past events. What is more, in so far as the various celestial bodies are moving, the place that the astronomer is looking at when he observes the remotely past phenomena is not even the place at which they happened - that place will now be somewhere else.

In passing, we make tacitly the assumption that words like here, there, now, then, ago have the same connotations when referring to huge distances and huge lapses of time as they have for short distances and brief durations. In other words we assume tacitly that the differences between a mile and billion miles, the difference between a year and a billion years - are purely quantitative rather than qualitative. May we properly assume that the very big is like the very small, only bigger. We have no necessary reason to make such assumptions even though they seem not to matter when comparing the everyday smaller with the everyday larger quantities in life. I do not propose to explore this line of questioning further but one does have to say that is there to be explored.

It is obviously the case that, short of truly revolutionary technical advance, we cannot know what is going on now at a site half a million light years away. So all our knowledge of the distant cosmos in particular (and strictly speaking of nearby parts too) rests upon 'seeing ghosts' - and this must be so in so far as the velocity of light, and other kindred radiations, is finite.

We now have to ask on what assumptions dare we suppose that the observation of distant things and events - long after the things have changed and long after the events have run their course (somewhere else!) - be relied upon as data in the study of those things and those events?

Those assumptions must include that the things and events observed gave rise to energy packages, that are structured in such a way as to tell us something meaningful about their origins, have survived huge lapses of time, while travelling huge distances to reach us, with those structures not corrupted to such an extent as to vitiate their relevance to the situations in which they originated .

This assumption is a bold one but it has to be made if any inferences whatsoever are to be drawn at all from the observations of distant long past phenomena - observations that are the staple of astronomy and cosmology.

To have to make this assumption of the astronomical while, implicitly or explicitly, denying it of the terrestrial seems like straining at a gnat but swallowing a camel. If we can observe a stellar explosion of half a million years ago (because of the durability of the structured energy packages the event emitted) then it seems to be not impossible for the structured energy packages emitted now by me in this chair, wearing these clothes, might survive for, say, fifty years and be seen as my ghost some years after my death (an event likely to occur before fifty years have elapsed - considering that I am in my eighth decade).

It may be objected that the stellar explosion emitted enormously more structured packages of energy than I am now emitting. True enough, but it would need to do so for its packaged energy to last for huge times and over huge distances - vastly more than the thirty years after my death when I might be 'seen' perhaps only a few feet away from where I am now.

Readers will be tempted to ask - of my hypothetical ghostly appearance a few years hence - how can this be possible, what mechanism, whereby it might happen, can we envisage?

Readers may, by now, be wondering whether this is all a donnish leg-pull or simply a sign of my dotage. That is their problem. What I have to insist upon is that it is simply not a valid argument to say that I do not understand how this could be the case therefore it cannot be the case. Readers would do well to remember the once-alleged impossibility of going to the east by sailing west ... the once-alleged impossibility that iron ships could float ... the once- alleged impossibility of ... whatever ...

As a chemist, I am reminded of a remark by John Dalton (the founder of modern atomic theory) in 1808 or thereabouts. He said something like "We might as well try to put a new planet into the Solar System as try to split even one atom". He was, in a way, quite right; by 1958 we had done both even though the planet in question was a very small earth-satellite, a sub-planet in the Solar System, Sputnik