Evaluating armorials (IV) – Grünenberg, the unfortunate armorist
The last post in the series on evaluating armorials is a case study of the famous Grünenberg armorial, in which different elements that have been discussed in the previous posts return.
There is one question that is seldom asked: How reliable is the studied armorial – or to rephrase it: Can we trust a single occurrence of a coat of arms? The answer would in any case be highly subjective and dependent on the armorial or other material studied. As an example, we can use the Grünenberg. An armorial of fine repute that is easily available and frequently cited. A kind of facsimile was published in 1875 and two recent editions are available. The manuscripts have been described by Christof Rolker and Bernd Konrad on this blog.
The Grünenberg is not an autograph compilation, but a book commissioned by Conrad Grünenberg, a citizen of Constance, from a commercial workshop. It is likely that Conrad Grünenberg himself was closely involved in the selection of the arms and the editing of the various parts, but the artwork was done by a professional painter, and much of the content was copied from older material. As such we may characterize Grünenberg as an enthusiastic amateur armorist.
Variations and proper coats of arms
As described in part II of this series, there are several types of variations, of which only the first two are serious. These are: (1) different main figures; and (2) inversions (or change) of colour, both of which would represent a different family or at best a different branch. Less serious variants are (3) reversions and flippings, with the colours (or quarters) being mirror images of the proper order; (4) wobbling, changes in the number of partitions of the field; and (5) fading of colours. The latter three variants are easily reconciled with the proper arms. That is, when the proper coat of arms has been decided.
In the end it is a personal choice that decides which blazon describes the ‘proper’ coat of arms. At best the family is still living, or there is a substantial concordance of descriptions among the most trustworthy sources. If not, one needs to assess which sources to trust, and which sources only contribute circular information. For a considerable number of arms and names, there are no seals or any other reliable references available, while for others the modern reader simply does not know their existence. Such name-arms combinations of ‘families’ can be said to be not verified. The term does not imply that the name-arms combination is not real, but simply that independent evidence of the connection was not found. Neither does it mean that no other instances of the arms are known. Among the 100 Germans noted as not verified in Grünenberg segment 23, there are 26 items which may be found in some of the armorials examined. Among the 26 items there are 6 cases of variation. One, an Argent-Or exchange, might be due to the condition of the manuscripts, another is a simple flipping of colour in a barry coat-of-arms. Three have variant blazons – like the Ochsenstein with different combinations of colours and design of the chief. Names can be misleading too, e.g. Vudberg for Neidberg, a cadet of the Austrian Stubenberg. Their quarters are reversed and unfinished in a later French armorial.
The reliability of an armorial (or a part thereof), can be defined as the proportion of proper coats of arms in it. In order to determine the reliability, one hase to take the following five steps:
1) remove any double entries, so that only one instance of each coat of arms remain in the population. Of the 906 arms evaluated (517 from segments 14-22 + 389 from segment 23 after removal of those not verified) only 17 families are represented more than once;
2) remove items that cannot be verified as representing real families (and of course reduce the total by the number of removed items). This excludes a large part of the Grünenberg and leaves two chapters with nearly 1000 items from the territory between Alsace in the west and Vienna in the east, and between Karlruhe in the north and Bern in the south.;
3) determine the proper coat;
4) determine the number of variant coats per type;
5) calculate the percentages (or frequencies) as in the example below.
The two groups of segments (14-22 and 23) apparently have different origins and ought to be examined separately. The first group has a total of 192 variants (thus mistakes) with 102 serious variants or 20% of the population of 517. The aggregate total of variants is 37% for the first group, but only 28% for the second group. For unknown reasons the frequency of less serious variants in the second group was halved.
The result of the analysis is rather overwhelming. If we knew only this occurrence of each of these 906 arms, we could be wrong for nearly every third coat of arms. More seriously, we would be gravely wrong every fifth time, if we trusted this source alone. If one prefers these results to be displayed in a reliability index (1 – %-variants), please chose any number between 0.63 and 0.8. How did Conrad Grünenberg end up with such a bad result?
One of the main reasons for the unreliability of the Grünenberg armorial as a source of family arms is that he relied on sources of uncertain provenance for his information on the coats of arms. Some of these must have had a considerable number of badly copied items. Though the artist or artisans probably had some experience with the painting of arms, he (or they) were not armorists and knew very little of the items they copied or interpreted from written instructions. This could easily add to the number of mistakes. Conrad Grünenberg may have known a little more of the families involved, but had probably only seen the arms in use by very few. So he too would have to rely on what was in the books and notes that he was able to borrow or buy. Finally, once a coat-of-arms was painted, there was little chance of revising it. In summary, without quality control, the customer unfortunately received a splendidly painted mess of family arms. On the positive side, most armorials appear to have a much lesser frequency of serious mistakes.
 Rudolf Gf. Stillfried-Alcantara & A.M. Hildebrand: Des Conrad Grünenbergs Wappenbuch. I-IV. Görlitz 1875. Michel Popoff: Armorial Grünenberg. 1: Édition critique de l’armorial de Conrad Grünenberg (1483). 2: fac-simile pubblicato a Görlitz nel 1875. Milano 2011, introduction (35 pp) by Michel Pastoureau. Steen Clemmensen: Conrad Grünenbergs Wappenbuch. Farum 2009, www.armorial.dk.
 http://heraldica.hypotheses.org/464 , 468, 504, and 678. The image is from page 17 of the München or BSB manuscript, Bayerische StaatsB ibliothek, Cgm. 145, available on http://daten.digitale-sammlungen.de/bsb00035320 .
 The primary manuscript is the Berliner or GStA version. See the introduction to Clemmensen Grünenberg, and the blogpost 464 in note 2.
 See Clemmensen : Evaluating armorials (II) – variant coats of arms.
 Neidberg, quartered arms, in GRU:2162 / 350n2, NLU:999 / 79v10. With Q2 (talbot / wolf rampant) only in ING:67 and SGH:168n1. NLU, Nancy, Bibl.Municipale, ms.1747, on: http://bmn-renaissance.nancy.fr/items/show/1240 .
 Grünenberg segments 14-22 (members of tournament societies) and segments 23-24 (residual of mainly people from Southern Germany, but with a sprinkling of men from Mecklenburg and other places), data from Clemmensen Grünenberg and Steen Clemmensen: Ordinary of Medieval armorials, Farum 2013, www.armorial.dk, described in http://heraldica.hypotheses.org/683.