Chasing a reference

This post being an account of trying to chase down a reference in a paper about dreaming. A paper which is now around 13 years old. A chasing down which soaked up rather more time than I was expecting.

Figure 1: The text

Figure 2: The reference

Figure 3: The abstract

The model

We want to say something about the extent to which dreams are derived from real life. To what extent can any particular fragment of dream be said to be derived from something which actually happened in real life? 

Most dreams are built of images and words which are derived from memory of one sort or another, memories which are derived, in the main, from senses internal, peripheral and external. So the raw material of dreams is drawn from real life, even if the bricks of raw material are assembled in a new way. But can we say when the bricks have been put back together properly?

Without getting into the detail of dreams, it turns out to be possible to analyse both them and the real life events from which they may have been derived in terms of a number of features or properties. In the present paper these properties are: characters, objects, actions, themes, emotions, and location. Then we can ask to what extent the properties of a dream fragment match those of a real life event or episode. Which gives us a way into our problem: is this or that dream fragment about, derived in a straightforward way from, this or that real life event?

These properties are not properly defined, but some support is provided by the exclusions – things which were not deemed to match – of table 2. And I have taken characters to be the people involved, objects to be other things of importance, actions to be the important actions, themes to be very short summaries of what the fragment or episode as a while is about, perhaps just a few words, emotions to be the dominant emotions that came with the fragment or episode and locations to be the particular places. So not just some London square, but Trafalgar Square. As it happens, the present paper puts of a lot of weight on location.

Time is not one of the properties used for matching. Which attracts some comment. First, the time of the dream is necessarily different from the time of the material from memory which is driving it. Second, the time of the dream could be a season or a festival, without specifying the particular date or year. So the dream could be a New Year’s Eve party, and the driving event could be the party of 1989-90. Third, my own dreams do not seem to involve time at all: no seasons, festivals, days of the week, months of the year or years. Not even time of day. Never mind the state of the moon – this despite my having tracked the state of the moon since the state of the plague, that is to say the beginning of 2020. Yesterday, for example, by late afternoon there was a fine crescent moon to the south.

In what follows I use the distinction between episodic and semantic memory – without much caring for this particular binary distinction, but it will serve for the present. Wikipedia provides a simple definition of these two sorts of memory at reference 6 which I gloss as follows: an episodic memory of pork soup is a record of some particular occasion in my life, an occasion which has a time and a place and which involved pork soup in some way or other. For example, that noticed at reference 5. The semantic memory of pork soup on the other hand is what can be said about pork soup in general. With each pork soup episode creating the opportunity for the brain to create a new episodic memory and to dust off, to update the semantic memory. Perhaps, for example, the pork involved gradually slides from being tenderloin to be being any kind of coarsely chopped pork meat. Or the volume of soup made on any one occasion gradually drifts up from two litres to three.

The story

In the course of reading Hobson’s short book on dreams at reference 1, I jumped to his later paper at reference 2. Where I read that only 20% of dreams draw on events in the dreamer’s life history – for which see Figure 1 above. Some time later, I remembered about this, and wanting to use it, thought that I better check it, eventually getting my own copy of the paper at his reference 58 above, my reference 3 below.

Off to a bad start, in that searching this paper for the string ‘20%’ fails to find anything.

But what we do have is 29 students, half male, half female, writing down their waking and dreaming doings for a period of 14 days. This was normal life in the real world, not dream laboratory doings. So we lose the control and the records available in a laboratory, but we also get out of that artificial world into the real world – which is better for present purposes. Inter alia, this generated 299 dream reports of which 194 contained candidate memories from life, 364 candidates in all.

Figure 4: Table 1

Candidates which were progressively whittled away by applying stricter and stricter criteria for what counts as a memory entry derived from episodic memory.

Whittling away was supported by experiential features: ‘… Each of the 297 remaining memory entries were scored as similar to a waking event on one or more of the following experiential features: characters, objects, actions, themes, emotions, and location. The 297 memory entries included a total of 973 scored features, with the frequency of occurrence for each feature shown in [his, not my] Figure 1…’.

My guess is that the Hobson 20% of reference 2 is derived by taking line E in the table above, with both reports and entries coming in at something less than 20%. In which location is the big driver: a dream memory entry is not deemed to be based on episodic memory unless, inter alia, it happens in the same location in both places. I think that this requirement is too strong. And furthermore that it is not well glossed by the words used at Figure 1 above.

While line J suggests that very few dream fragments reproduce a real life episode in all its essentials. The authors suggest that this is consistent with poor or no access to episodic memory, particularly during REM sleep. Which I am content with: my own dream diary – now covering getting on for six months and including getting on for 200 entries – does include any such dream or dream fragment. Drawing on a real life episode, very possibly; reproducing one in all its essentials, probably not. For a first report from this diary see reference 8.

By way of example, a recent dream of my own involved something very like pork soup, the soup which was the subject of the report at reference 5. I believe the fact that I had made pork soup in the recent past, put that soup near the top of the heap when it came to the sleeping brain selecting stuff to add into a dream. To that extent, that bit of my dream is the result of a recent episode in my life. But pork soup apart, it does not make much other use of that episode, for example the occasion, the cooker or the saucepan involved. The bit of dream might be no more than a vague image of soup in a saucepan linked with a label saying ‘pork soup’. 

So my bit of dream need not involve any memory of the recent pork soup episode at all. Perhaps, instead, the making of the soup reinforced the semantic memory of same and put it near the top of the heap from which the dreaming brain takes its ingredients. 

Or even this may be too strong and the soup’s appearance in the dream and its recent appearance on our table may simply be a coincidence – coincidences which are going to happen from time to time, but presumably only in a small percentage of dream features. And maybe one notices these coincidences, making more of them than one should. In which, I, in common with plenty of other people, find it far more satisfactory to have a cause rather than a random event. We like the idea that the fact of the soup being on the table put that same soup into the dream. We are comfortable with that notion and stick with it.

I should add that it took me a while to register that this paper was as much about where in memory the dream came from, as about where those memories came from in real life. Many sleep and memory researchers, in particular these ones, are interested in the way that sleeping in general and dreaming in particular interact with memory consolidation, and in that context the difference between episodic and semantic memory is important. 

But I think that Hobson should, nevertheless, have been a bit more careful with his 20%.

Other matters

The dreams that Hobson reports in his book (reference 1) are quite different from my dreams, which may be an artefact of their recording, but it may also reflect differences between us. In any event, I suspect that dreams vary a good deal between people, between different kinds of people and across time. It may also be that that this variation does not change the present fact much, that few dreams are based on episodic memories, although they may be prompted by real life events. Against which background, I note that the present sample is a rather small sample of rather special young people. It would be interesting to know whether anything like this experiment has been repeated with larger or at least different samples.

I am reminded of the Suedfeld paper at reference 7, already noticed at reference 4. Defining exactly what we mean by stopping smoking – if we stop short of completely stopping with no relapses at all – is more complicated than one might think. As is defining exactly what we mean by a dream not drawing on episodes in real life.

Conclusions

While one might argue about the 20% figure given above, it does seem to be the case that few of the bricks of which dreams are made are simply chunks of real life. They are processed real life, extracts from real life – rather as saucisson sec is processed pig, not connecting in a simple, visual way to a live pig.

But I also think that Hobson might usefully have been a bit more careful in using this figure. He might, for example, have provided a bit more background.

References

Reference 1: Dreaming: An introduction to the science of sleep – J. Allan Hobson – 2002.

Reference 2: REM sleep and dreaming: towards a theory of protoconsciousness – J. Allan Hobson – 2009. A paper written when Hobson was in his mid seventies, a little older than I am now.

Reference 3: Dreaming and episodic memory: a functional dissociation – Fosse, M. J., Fosse, R., Hobson, J. A. & Stickgold, R. J. – 2003.

Reference 4: https://psmv5.blogspot.com/2022/11/sleepy-sickness.html. A previous spin-off from Hobson’s book.

Reference 5: https://psmv5.blogspot.com/2022/12/pork-soup-redux.html. 

Reference 6: https://en.wikipedia.org/wiki/Semantic_memory. 

Reference 7: Restricted Environmental Stimulation and Smoking Cessation: A 15-Year Progress Report - Peter Suedfeld – 1990.

Reference 8: https://psmv5.blogspot.com/2022/10/dream-diary-first-report.html.