Primary v Secondary Sources

[PhilB]

Some critics have been very sniffy about the use of so called secondary sources. As long as the source fairly reflects the actuality, I can`t see that it matters, but then I`m no historian. It almost seems to be a case of "Primary source good, secondary source bad" and I would have thought that a primary source is just as likely to be misleading. Any historians out there who`ll explain the differentials?

[truthergw]

Whatever the source, the person quoting from it is making a selection. A secondary source is then a selection from a selection. This increases the danger that the quote is in fact not a fair representation.

[Muerrisch]

Depends on the definitions.

If I quote from a War Diary and acknowledge it, that's a primary source .... it might be wrong, it might be right, but it is primary.

If I quote from a regimental history that purports to quote from the War Diary, that is secondary.

I only ever use secondary in my books and articles where nothing else availeth, and even then with a big caveat, and preferably from two sources.

Never mind historians, anyone with a scientific training would always go for primary.

Sometimes this costs a lot of time and money in traveliing. So be it.

[dycer]

I'd agree primary sources can be doubtful.

An obvious primary source is a War Diary.Assuming they were written daily,I would suggest,in some areas,they will not be accurate.

As an aside,in my TA days,I was a Company Signaller,The Signals Bod,had the bright idea,during one exercise we would keep a written log record of all radio traffic.My verbatim one lasted about 5 minutes,and the rest was completed from memory.You will appreciate that none of Company logs agreed,when reviewed at the end of the exercise e.g. time of HQ message,content,etc.

George

[charlesmessenger]

I certainly agree with Grumpy that one primary source on its own is not necessarily solid evidence and does need confirming from other primary sources. The problems with employing merely secondary sources are twofold:

a. If they draw on primary sources, it is often the writer's interpretation of them and this is not necessarily the correct one.

b. There is a danger of errors of fact being regurgitated by one book from another.

On the other hand, there is the danger of misquoting primary sources or taking them out of context so as to fit in with the writer's thesis - Denis Winter's Haig's Command is a now classic example of this. Some historians are more reliable than others and there is certainly no harm on drawing on the work of one who is generally factually correct and whose views you respect.

Charles M

[Alan Tucker]

They key thing for any historian is to treat all sources - primary or secondary - with a healthy degree of scepticism and always apply certain basic questions about reliability before use.

[HarryBettsMCDCM]

In actuality are not War Diaries secondary Sources; albiet contemporary ones,the Primary source would {Have} been those persons there @ the time,by the time it is written down it surely becomes a secondary source??

[4thGordons]

Funnily enough I was looking a just this question today....

HERE is a discussion of primary sources

and HERE is the same people's (UIUC) definition of a secondary source

Chris

[Robert Dunlop]

All sources have to be approached with caution. Primary sources have many advantages, but can also cause significant problems if not understood or interpreted appropriately. By 'appropriately', I do not mean 'interpreted to give the correct one-and-only interpretation'. I mean that some thought is given to what the limitations are.

The concern about secondary sources reflects the growing number of primary sources that are now available and the significant growth in people who are researching the Great War in a more rigorous way. The latter is helped by the rise in numbers of academic departments specialising in military history, even specialising in the Great War. Research is no longer the domain of the war colleges and a few individuals. Several Forum members have published books, theses or dissertations. Others have a deep interest and have read widely, perhaps with an academic background in another domain but often not, which does not detract at all from their contributions.

There are several examples of the problems caused by secondary sources. I am reading Shaw Sparrow's book 'The Fifth Army in March 1918', published in 1923. It is a secondary account of Gough's Fifth Army during Operation Michael. A classic example is mention that, to paraphrase, because a British unit lost heavily then their German counterparts who attacked them must have lost heavily too. This attribution is not unique to secondary sources, however. It is often seen in primary accounts as well.

More serious is the effect of incorrect conclusions in secondary sources, especially when these get passed on as the definitive view. Mention has been made of Denis Winter's work. Another example is the works of Prior and Wilson. I have touched on the specifics elsewhere. Some of their more important conclusions have been quoted widely in other works but do not bear close scrutiny when the primary sources, from which they often quote (selectively), are examined in detail.

The most significant problem, IMHO, has been the dearth of information about the other perspectives in the war. The German and French sources are examples, but it has been great to see the emergence of information from Turkish and Romanian sources, to name just a few, on this Forum. Over the years, there have been writers who included these perspectives, but in recent times the contribution of authors like Bruce Gudmundsson, Jack Sheldon, Annika Mombauer, Robert Foley, Antulio Ecchevarria and many others has totally transformed the standard by which we should review accounts of the war. Not forgetting the contributors to this Forum as well, many of whom study primary/secondary sources in other languages, and/or bring a very different perspective to bear on many issues. Thank goodness for diversity!

Getting back to primary sources, contributors to this thread have already exposed some of the issues. As long as humans record events, there will always be biases in what is perceived, what is remembered, and what is recorded. Personal accounts are wonderful to read, especially if we want to learn about what it must have been like for a relative. But these accounts often reveal the very very narrow perspective of an individual in the midst of a battle. In the past, a single German soldier's recollection might be included as the view of the other side of the hill (no prizes for guessing which source is most widely quoted in this respect). The same thing has been done with British accounts too. Spears' observation about the terrible state of some French cavalry horses that he saw is a good example, where all French cavalry are now judged by this source. Authors like Lyn Macdonald, Martin Middlebrook, Peter Hart and Bryn Hammond have really helped us to understand why it is necessary to study multiple personal anecdotal accounts if we are to gain a broader understanding.

And there is still so much to learn, so much more to study. Many primary sources have barely been touched.

Therefore it is not a case of primary sources good, secondary sources bad IMHO. All sources have strengths and weaknesses. The key is to understand this is the case, even if we don't fully understand what the specific strengths and weaknesses are. The corollary is not to place too much reliance on any source, and to retain a healthy scepticism. A challenge is, in the context of the GWF, to express this scepticism in a manner that does not criticise or imply criticism of the individual on the receiving end. Harder still is to receive negative or questioning comments about one's work without feeling this as an attack on one's self. It is very noticeable that GWF Forum members cannot post thoughts and conclusions without expecting one or more replies from experts in that particuarly aspect of the Great War. I love this aspect of the GWF. Great to have this diversity and it has certainly driven out many new insights that I have thoroughly appreciated.

Robert

[Anthony Gorst]

Robert

Fully agree - all sources have their flaws, whether primary or secondary, but sticking with primary the key is to consult as wide a variety of sources as are available from contemporaneous accounts to memoirs to oral histories etc. Then stick it in the mixer and see what pops out (to mix my metaphors).......

Edit for further thoughts - I have have some reservations about oral history and it's reliability (based on experience on a non WW1 project, but it can be useful for 'local colour'

[Pete1052]

One of the trends of the last 20 years in the study of the American Civil War has been to place less reliance on primary sources that were written years or decades after the war. Men's memories of events change with time and often memoirs are self-serving. Things written after the fact by participants in controversial events often have a particular axe to grind.

[truthergw]

Hi Pete. I believe the time scale to be fairly irrelevant here. Contemporary accounts are just as liable to be biased as those recorded many years later. We have to differentiate between facts and interpretations of those facts. If in doubt, we need to look at the author, try to fathom his mind set then allow for it. George McLellan's and Abraham Lincoln's accounts of the Civil War will differ by a wide margin. Contemporary, primary sources both but they will need to be filtered by our knowledge of the men and their time.

[Rockdoc]

I would not call memoirs a primary source. They are bound to be both selective and subjective and, therefore, have to be corroborated from other sources. Memory plays very strange tricks and it's well known that witnesses to a crime will almost certainly tell wildly different versions of the same event, even assuming that the author of a memoir isn't guilty of self-aggrandisement! Anything that is not a simple record of events, such as a War Diary, should be treated as suspect until verified and it's become fairly clear to me recently that some War Diaries were written up well after the events they record.

The other point to bear in mind is not to be swayed by the source of the information. Just because it was written by an official body doesn't mean no twist was put on the facts.

Keith

[thesamsonsed@gmail.com]

I would not call memoirs a primary source.

I would definitely call memoirs a primary source as they are the author's thoughts - they may have been reworked from the original record the author had, but they are his/her personal recollections (purposefully tainted or not).

Just to complicate matters(?), I see all sources as potentially being primary - it depends on your use of the document. In my work on memory of the EA campaign of 14-18, I have made use of novels and newspapers to understand how people thought about issues at a given time. Although these are generally regarded as secondary (if historical at all in the case of novels), I'm using them as primary. The main thing is to treat all sources (primary and secondary) with suspicion - they were all written for a specific purpose and therefore are likely to give a different take on the same event. Our role as historians (traditional, family, GW etc) is to treat them all with caution, test them against what others have written (and see how they fit into the jigsaw).

It's also helpful to remember that we don't, and won't likely, have access to all the sources on a given event which is why we keep having new publications come out on the same thing (think of all the biographies on the WW1 Generals). As people discover new sources or revisit secondary ones written near the event's occurance, so new interpretations are made. Rather than get bogged down with definitions of what is primary/secondary, as others in the discussion have said, treat them all with care.

Best wishes

Anne

[Greg]

And-just to add further confusion- a 'secondary' source can also be a primary source. So for, example, a school textbook written in say the 1920s may not tell us a great deal about the history of its subject but may tell us a great deal about how people at the time, especially the authors of school textbooks - thought about it.

In studies of the Great War the problem is that there is a vast literature much of which draws upon earlier works- some of which were written at a time when there were limited source materials available and others which were not written from as scholarly perspective .

Inevitably however great the number of primary sources consulted (even allowing for their bias) the way in which they are assembled and placed in context by an author will have a great effect on wht they say to the reader.

Greg

[Desdichado]

Depends on the definitions.

If I quote from a War Diary and acknowledge it, that's a primary source .... it might be wrong, it might be right, but it is primary.

If I quote from a regimental history that purports to quote from the War Diary, that is secondary.

I only ever use secondary in my books and articles where nothing else availeth, and even then with a big caveat, and preferably from two sources.

Never mind historians, anyone with a scientific training would always go for primary.

Sometimes this costs a lot of time and money in traveliing. So be it.

My old history professor's maxim was "primary sources always get the veto."

[dycer]

The ultimate primary source now left,for experiences,are the few WW1 survivors,still alive.

But with the greatest of respect,to them,are we sure their recollections are now accurate?

George

[truthergw]

I would not call memoirs a primary source...............

Keith

Sorry. Although everything you say makes good sense, memoirs are primary as long as they are not derived from another source.

There seems to be some confusion as to what constitutes a primary source. A first hand description of an actual event. The minutes of a meeting would be a primary source. A person quoting a letter from the trenches to describe what a battle was like would be a secondary source. There is no implication of truth or accuracy in either. Simply an indication of the immediacy of the description.

[Rockdoc]

I see I opened a can of worms! Taking a slightly different tack to try and explain why I made my statement the way I did, think of a scientific or engineering higher-degree thesis. Although the thesis is written by the person doing the research it does not contain raw results, only the author's summaries of the experiments undertaken. That, to me, makes the thesis a secondary source because it cannot be taken at face value unless the results are included as an Appendix, which practically never happens. The primary source is the lab note-book that contains the data as it was recorded at the time. I classify a memoir in the same way as I classify a thesis: something that is an excellent starting point but which should not be treated as Gospel.

Yes, I'm cynical. I have read theses where deliberate attempts have been made to present data that actually contradicts the hypothesis being proposed in a favourable light. I have read one, which was rejected by the examiners I'm glad to say, where much of the data had obviously been faked. Those are exceptions - a very small percentage of theses submitted - but if the reader does not have the experience to spot the deliberate mistakes and downright deceptions then a memoir must be treated as a secondary source and, as I wrote before, corroborated by at least one other source.

Keith

[4thGordons]

I see I opened a can of worms! Taking a slightly different tack to try and explain why I made my statement the way I did, think of a scientific or engineering higher-degree thesis. Although the thesis is written by the person doing the research it does not contain raw results, only the author's summaries of the experiments undertaken. That, to me, makes the thesis a secondary source because it cannot be taken at face value unless the results are included as an Appendix, which practically never happens. The primary source is the lab note-book that contains the data as it was recorded at the time. I classify a memoir in the same way as I classify a thesis: something that is an excellent starting point but which should not be treated as Gospel.

Yes, I'm cynical. I have read theses where deliberate attempts have been made to present data that actually contradicts the hypothesis being proposed in a favourable light. I have read one, which was rejected by the examiners I'm glad to say, where much of the data had obviously been faked. Those are exceptions - a very small percentage of theses submitted - but if the reader does not have the experience to spot the deliberate mistakes and downright deceptions then a memoir must be treated as a secondary source and, as I wrote before, corroborated by at least one other source.

Keith

Doesn't it rather depend what you are researching? To use your example - if my interest were the student (the person doing the research), then his/her work (the thesis) is a valuable PRIMARY SOURCE into how they were thinking at the time, which theories they then found most appealing etc. Or, if my interest was in the educational establishment or education policy, the thesis may be a primary source as an example of contemporary requirements for a degree qualification etc etc.

So for me a memoire is clearly a primary source if you are interested in the author (this remains true even when the author is attempting post facto rationalisations)

If you are interested in the events described I would argue it is still a primary source but one to be treated with circumspection (as you say, needing corroboration)- nothing makes primary sources intrinsically reliable as accounts of events (ie needing no corroboration), but I think they are a reliable indicator of one participant's view of the events.

Chris

[Rockdoc]

I agree with your rationale 100%, Chris. As an engineer I suppose I tend to think much more in terms of the tests and results than the people doing that work. I'm sure if you were researching the person whose PhD was rejected by the examiners it would be of great interest to discover that he faked the results because he did not believe the results he got from actually doing the tests - and he had done extensive work - were good enough to satisfy his supervisor. That would be quite enlightening as to the relationship between the two as well as to the amount and thoroughness of any proof-reading done by the supervisor before the thesis was submitted.

Keith

[Pete1052]

I believe the time scale to be fairly irrelevant here. Contemporary accounts are just as liable to be biased as those recorded many years later. We have to differentiate between facts and interpretations of those facts. If in doubt, we need to look at the author, try to fathom his mind set then allow for it.

A lot depends on what precisely is being recalled. As a general rule descriptions of tactical detail involving the date, time of day, location, weather, etc., become less accurate the later after the event that they were written. Without referring back to our checkbooks many of us would have a hard time recalling what day it was when we last paid the electric or water bills, much less when we did so 20 or 30 years ago. That having been said, some military men who publish memoirs use detailed notes, correspondence, diaries and so forth when they write their books, so each source has to be evaluated on its own merits.

To cite a classic example, some Civil War recollections written years after the fact by eyewitnesses have proven to be very problematic when they discuss what they claim Confederate General James Longstreet did or did not do at the battle of Gettysburg. If the topic is highly charged and controversial it is best to rely on contemporaneous records and not on what someone says was the case 30 years later.

[Robert Dunlop]

...a scientific or engineering higher-degree thesis

Keith, thanks for providing your frame of reference. Scientific journal articles are another example. In some cases the raw data is published. Mostly, the analyses are reported, along with the conclusions. So long as the authors were directly involved, these papers are often regarded as primary sources. Other papers, editorials and reviews that quote the primary sources are regarded as secondary sources, as you know. Your point, if I understand correctly, is that the 'primary' sources are often 'secondary' because they only reference the original data. Thus statistical analyses and conclusions are 'secondary', even if carried out by the the team who recorded the findings. This is an alternative definition of 'primary' and 'secondary' sources, which is different from, but not better/worse than, the definitions referenced earlier in this thread.

It is interesting, however, to consider the issue of raw data in historical studies. The recording of this data is much more complex. It is as if the principal investigators and researchers had no instruments for generating data, didn't realise they were actually performing a study, or were so caught up in/overwhelmed by the process of the study that they could not record anything at the time. Some data was recorded at the time, such as the timed messages that are recorded in battle diaries or on the original message forms. But even these contemporaneous 'data' sources are nothing like the printout of an automated analyser for example. Much of the 'data' is equivalent to a researcher sitting down after the project, sometimes years later, and trying to recall what the results were.

Robert

[PhilB]

It is unfortunate that "Primary" has connotations of reliability and integrity higher than "Secondary". The only documents I would be happy to call primary are those written before an event such as battle plans and intentions. Anything written after the event must be suspect for the reasons given in previous posts.

[Rockdoc]

Two things are emerging from this thread, I believe, and both would have a significant effect on how a source is perceived.

The first is the background of the researcher. I'm an engineer, as I wrote before, and spent most of my working life in laboratories and in on-site research programmes. My experience in those fields makes me distrustful of any conclusion that I cannot verify for myself and it's likely that my attitudes were shaped to a large degree by my first job where we always took three copies of the logged data (1 original and 2 carbons, which shows my age!). The company kept the original, the client got a copy and the third went to the consulting engineers so that our report could be put under very close scrutiny. Given that the tests were to complete large capital projects and millions rode on them it's no surprise but that philosophy has been at the back of my mind ever since. In researching the AA Sections in Salonika, my instinct is to look at their placement and operation both individually and as a coherent defensive structure. It may be that someone with a history or sociology background would approach the same topic from the direction of the men involved and how their experiences varied from Section to Section. I wouldn't know where to begin such a project.

The second is closely related to the first: the end that the research is focused on. As has been pointed out, sources I would regard as secondary can be primary under other spotlights.The question of whether Pte Fred Bloggs shot down Manfred von Huffenpuff with a Mk IV pea-shooter on a given date needs alternative verification, wherever it's recorded but, even if the event is proven false, the ground is probably still fertile for a different researcher to examine the situations that might make troops make false claims and then Bloggs' statements would be primary.

Ultimately, I think the separation into a primary or secondary source can be defined by the degree to which each needs corroboration before the information it contains can be fully trusted. That, as has been clearly shown in this thread, depends quite strongly on the nature of the research. However, the influence of the researcher is a big factor because we've seen huge differences in approach.

All in all, another fascinating thread.

Keith