Tuesday, November 16, 2010

Ming Inspiration (4)

The last couple of posts in this thread have delved a bit more deeply into classical Chinese corner leg tables, with a look both at the many varieties of the table form, and a detailed look at the joinery methods employed at the junction of leg to the table frame. Today I want to turn my attention to a particular Chinese table which is really quite novel and forms the basis in many ways for the dining table I am to build shortly. I thought it worth posting up a picture of that side table again:

This table features very clean lines with no structural support assistance from braces, spandrels, or stretchers - only the joinery at the corners maintains the integrity of the construction. Though the top of the table is not clearly visible (and I have no other pictures unfortunately), it can be seen that the top is a seamless wooden surface, with no telltale spaces between frame and panel, as one might expect, such as can be seen in the following table:

Since there is no gap for a panel to move around in, and given the clean transition to the sides of the table from the top, one might conclude that the top of this table is made from a thick slab of wood. This would be an unusual construction for a corner leg table, as most Chinese tables which employ a thick slab of wood for the top are of a different variety. One type is called a jiaoji'an (架几案), a table formed by a pair of pedestals with a plank (or frame and panel) spanning between them:

Not too many of these tables have survived to the present, unfortunately. It seems that during the 20th century cultural revolution in China, when anything remotely smacking of imperial or high class or even beautiful was considered corrupt and was often destroyed or disfigured. They called that 'a great leap forward'. Also, when people are starving and cold, the thick heavy tops of these tables often served nicely as excellent heating fuel. Sad but true, and if I had to choose between freezing to death and cutting up a masterpiece, well, it would be a hard choice.

The other type of table to have a thick plank top is a variety of recessed leg table - here's one that currently sits in the Palace Museum, Beijing:

This table is 343.5cm long (about 135.25") and the top plank is a full 8cm (3") thick. Gustav Ecke's book shows the table, but due to it's massive size was only able to present a corner view of the piece, and further details about the construction are omitted:

Tables with thick, wide solid slab tops are prone to several problems, and it is due to these problems that I generally avoid such construction in my own work:

  1. large thick slabs of wood typically feature a large percentage of tangential (flat) grain wood, which means they tend to move a lot in service. They also tend to warp, crack, and check over time. Inspection of old tables in museums will demonstrate that pieces constructed out of thick and wide stock suffer quite commonly from warped and cracked surfaces. Proper frame and panel construction is proven to better deal with wood's propensity to move over time, most especially in climates with pronounced seasonal swings in ambient humidity.
  2. slab tops are heavy and thus place a burden not only upon the framework of the table, but also upon anyone who wishes to move the piece of furniture at any point.
  3. heavy slab tops are prone to sagging over time from their own weight, with the practical span limit thought to be in the vicinity of 8'.
Point 2 in particular has significant repercussions over time. In the case of a table which may be moved about from time to time and placed on ground surfaces which may not be co-planer with the set up location of the table, the mass up high in the top-heavy table creates conditions of lateral and vertical loading which impose severe demands upon the structural joinery. Over time and much repetition these pressures could break down even the most sophisticated forms of joinery. Placing heavy items upon such a table, or sitting upon such a table as some people might do at times, only exacerbates the problem.

One might think that the solution would be to scale up the support framework accordingly, and this does work within limits, however that solution engenders a clunkiness to the appearance, and increases the weight of the table even further into a zone of unwieldiness. It's easy to build something massive and strong, but much more thought is involved if one wishes to build a piece that is light and strong.

In the case of a dining table there are also requirements for space below the table so that the diners may have room for their legs. Since the tops of dining tables sit at a maximum height of 30~31", and given the need for 24" of vertical clearance for the legs of the diners, one really has at best about 6" of room for the top and structural framework of the table. Of course, one can design a rectangular table so that the edges of the table do not have thick support rails, however in the case of a massive top the problems of sagging and warpage are ever-present, and finding a way to support that heavy top without a perimeter frame presents other challenges.

Another solution to creating a massive thick top is to fake it by the use of veneer. A plywood torsion box with veneered paneling could be made to look quite convincing and be very flat and strong, though in practice I dare say that the people doing veneer often seem unaware of how a solid plank's grain should look, relating top surface to ends and sides, and even worse, a large number of people who might buy tables are oblivious to how natural wood actually looks. So, people who veneer stuff 'unnaturally' often get away with it, as it were, since a significant percentage of the buying public cannot tell, in my estimation, whether something is solid wood or not. Even I'm fooled sometimes, especially by such things as chair seats which appear to have been formed from solid material - then I find that the veneer is actually wrapped around the seat's curvilinear surfaces. Of course, taking time to look at the underside and/or backside of a furniture piece is often a quick way to determine how it is made. I wish more furniture buyers did that, but we are culture heavily fixated upon the superficial and the frontal appearance of things and that does 95% of the selling.

So be it, however I will not be doing any veneering - though there undoubtedly is much art and skill in that arena of wood working, I believe in the virtues of solid wood despite its challenges and apparent 'limitations'.

With solid plank tops, the limitations are legion, so I tend to avoid that form of construction. The 3+ meter long table pictured above, now located in the Palace Museum, solves the problem of top-heavy construction by the recourse of hollowing out significant portions of the plank's underside. Not perhaps the most elegant or quickly-achieved solution, but one that seems to have stood up over time. That top lies flat and true and has not sagged.

Another approach to the problem is to start with a table top half the thickness you wish it to appear, and then attach it to a thicker perimeter framework, as in the following example from Ecke's book (plate #60), the "Psaltery Table":

The frame is not curved on the side as it appears in the photo - that's because the page in the book was curved and I didn't want to strain the spine of the book. Again, forgive my clumsy photography. As you can see from a view of the inside of the left leg assembly, the frame is about twice as thick as the panel. This is a table with some quite elegant corner spandrels I might point out. A "Psaltery Table", BTW, is a type of table used to rest a stringed instrument of the harp or zither family (a Japanese koto being one of those class of instruments).

The joining of the top to the sides in the Psaltery Table is accomplished by way of mitering the abutment between top and frame along it's length and employing stubby male dovetail tenons - as shown in the next illustration:

This technique of construction diminishes the weight of the top by some 40% and still provides a solid working surface.

Back to the table I'm intrigued with. It looks like it has a thick slab top, with only the corner joints to keep it standing. Several tables of this form and solid top construction exist in fact, but as mentioned earlier, the top heavy nature of these pieces makes the corner joins vulnerable to failure. It is thought that many tables from the late Ming of this construction developed loose joinery over time and were rebuilt with added-on stretchers. Few tables seem to have survived intact with such construction. Another point is that Chinese craftsman in the day would have been well aware of the structural problems with corner-leg construction, since they probably also repaired broken furniture from time to time. Various solutions were tried, and those that survive reflect at least a portion of the successes in that regard, and none of the failures. If it didn't stand up over time, it doesn't exist today, except possibly in pieces or in modified form.

So, the table I showed at the top of this page seems like quite an anomaly in several respects- and if you take a peek underneath you would be surprised to see that there is clear indication of frame and panel construction due to the presence of several battens running transversally across the underside of the top. Curators were puzzled too, as the top's appearance was not suggestive of a frame and panel, and so they decided to take the table apart to see what was going on. The top itself was held to the base structure by a pair of stub tenons, and a firm knock with a mallet broke the starch glue bond so the top could be removed. Somewhat unusually, one of those leg stub tenons was 'L' form in plan and was stepped. The aprons attached to the leg with half-lap mitered tenons and a thick tenon:

Note the small tongue on the end of the mitered apron piece, which fits into a corresponding groove on the leg.

Further knock down of the top revealed that the ends with the up-turned flanges (termed 'everted' flanges) were of one piece and have three narrow mortises corresponding to the top panel's three tenons:

Fully separated, an inner complexity reveals itself - note the dovetail tenons on the end of the frame's end member and the mitered returns below those tenons. Assembled, the flange member and the frame members sandwich the panel and are pinned and locked together with the above-mentioned protruding stepped tenons on the leg. The panel in turn sits upon a wide rebate on the frame.

But wait, there's more! The panel appears to be held conventionally, in terms of Chinese classical furniture construction techniques at least, by three transverse braces. These braces, or battens, have a male dovetail on their top surface which engages with dovetail mortises on the underside of the top - a construction which is intended to inhibit any propensity of the top panel to warp. However, in this table, the construction is taken a step further: as the long frame members are separated from the panel, further dovetail mortises are revealed, cut into the underside of the panel. These appear about every 9" along the edge of the panel:

When fully removed, the counterpart wedge-shaped dovetail tenons are revealed, these being carved from material left in the wide rebated areas of the long frame member. The dovetailed tenons were precisely cut, and it was clear to the restorers that the table top had never before been disassembled. Traces of a thin reddish lacquer were found inside the joints, evidently used for supplementary adhesion.

I'll quote now from the restoration account:

"When the table is reassembled, the long frame members slide neatly back into the dovetail mortises cut into the underside of the 1/2" thick panel. With the close spacing of the dovetail mortise-and-tenon joins, the long mitered join, where the edge of the panel meets the frame, is drawn and locked together to prevent any warping - almost four hundred years later this panel remains completely flat. The end frame members, cut from timbers large enough to shape the "smooth and rounded" raised everted flanges, are then slipped onto the tenons of the panel and into the large, complex mitered tenons extending from each long frame member. It is a remarkable example of design and execution, and amazingly, when all the elements are fully assembled not a hint is seen of the complexities within. Even the series of dovetail wedge-form tenons were intentionally cut back 1/4" from the inside of the frame, leaving no visible trace when fully assembled."

This table is without question a masterpiece of design and execution, a design proved by 400 years of survival structurally intact. Here's a picture which may make the genius of the construction a little more apparent if my above descriptions haven't made it clear:

The question is then posed, having taken in this work of a master who was clearly able to synthesize beauty with both form and function and craft a solution to a long-known structural problem with such incredible sophistication: was it practical to have put so much effort into such a design?" Many today might not hesitate to say 'no', since, in our modern strain of logic, time is money after all (or, put another way, money is what you are exchanging your life for, yes?).

Another view is noted by Craig Clunas, in his work Superfluous Things: Material Culture and Social Status in Early Modern China (1991) , wherein is quoted a Chinese philosopher writing on the moral nature of things, a certain fellow by the name of Wen Zhenheng (whose writings were compiled in the years 1615~21):

"In making utensils the men of old valued utility without sparing expense; thus their manufactures were extremely well prepared, unlike the slap-dash attitude of the men of later times...they delighted in refined elegance and did not vainly add inscriptions and value only signatures"

These sentiments are echoed in respect to the "vain addition of inscriptions", centuries later by Sōetsu Yanagi's work The Unknown Craftsman. There's something bittersweet, even ironic I suppose about someone in 1615, alive at the peak of the craftsmanship for the Ming period, complaining that the work of men of later times was slap-dash - kinda puts a lot of today's work into stark contrast, doesn't it?

The best and most fitting tribute to the master who designed and built this table is the fact of it's superb condition four centuries after construction. For me, this is the finest sort of work - craftsmanship without ostentation, and I strive to emulate the table described above, in the hope that what I make might live on four centuries or longer. Of course, a dining table carries with it a few different challenges as well, so I will be trying to stand upon the shoulders of the unknown maker of that table and take the lessons to be gleaned there and bring them forward. In the next post or two, I will begin exploring the design of this dining table and present to you the solutions I have come up with, all the while adhering to the rules/standards of classic Chinese furniture-making practice. I also want to delight in refined elegance with this dining table design, and fortunately I have found a client for this table who very much appreciates that exact attribute. I'll do my best to meet that high standard.

See you next time and thanks so much for visiting the Carpentry Way today.


  1. Amazing piece! How is the grain on the top/sides/ends arranged to give the look of a single continuous slab?


  2. Dan,

    the grain on the sides is in the same alignment as the grain in the top panel, and the wood is not highly figured, so it is not too hard to get them to match well where they meat at the edge. For the end grain portion, standard practice is to cover it up, through the use of cleated or 'breadboard' ends, or by way of using an everted flange end piece, as seen in the posting above on the long table in the Palace Museum. Covering the end grain is a means of slowing moisture exchange at the end grain, and cleating the plank can help control/reduce any propensity the material might have to cup across the grain.

    So, in solid wood, good practice means covering that end grain, so one would expect those pieces attached at the end of the plank to have grain running lengthwise. The only time you would see end grain at the end of a plank that would be the consequence of using a pure slab type of table top.


  3. Hi Chris,
    I am a little unclear on how the expansion of the "panel" composing the table top is handled. Do the mitered edge joints open and close with seasonal changes? The ends, or breadboards, seem to acting as as frame for the panel, so they will not be changing in length to compensate for the panel movement. Maybe I'm missing something. As always, I love your work and appreciate the effort you put into this blog.

    Harlan Barnhart

  4. Hi Harlan,

    on that table the mitered edge joins would be the places where seasonal movement would be accommodated. The edge of the board and the frame meet along that miter. You're not missing anything.

    That long miter cannot come to a sharp arris, as that would make the edge of the table vulnerable, and must be chamfered. As the type of wood used does not move much, the junction at the edge can actually conceal the movement if it is configured right. If this remains a mystery as to how exactly this is accomplished, then I would ask that you wait to see how I detail that part of the table I am making.

    As I haven't seen the original up close and in person, and only have a few photographs to go off of, I have to speculate upon some of the finer details, like the edge miter. Hopefully my head is screwed on right and the solution I see will prove to work well.



All comments are moderated, so if you're planning to spam this, know now that your clicking and pasting is in vain. I do read the comments before posting, so your mission is doomed from the outset. All this time and effort trying to put your inane spam onto blogs -- is this how you want to spend your time on earth?

Please do me the courtesy of appending your name to your comment, even if posting under the 'anonymous' option. No name = deleted.

Comments NOT accepted include:

-those containing links unrelated to blog content
-spam of any kind, or ham for that matter
-did I mention that attempted spam postings will be non-starters?