兽研1. 引言
如今，但凡有身份的人，谁都无法忍受被一只杂种狗跟随
这句名言引自19世纪多产作家威廉·戈登·斯托布尔斯1877年的作品，并由里特沃[1]引用，为本文主题提供了完美的序幕。维多利亚时代，英国贵族和中产阶级城市市民中休闲养犬的热潮空前高涨[2]，这一风尚迅速被工业化世界各处效仿和采纳[3,4]。我们认为，如果不了解现代犬种概念是如何以及为何产生的，就不可能解决当今犬主在根据周围社会的期望对犬进行最优且合乎伦理的调整方面所面临的许多问题。犬主和繁育者只能与那些经过过去150年纯种犬繁育有机发展所产生的犬只打交道，这一发展不仅塑造了犬的外貌，也定义了犬种的行为特征。
重要的是，维多利亚时代伴侣犬饲养的风尚聚焦于一项基本上是全新的“发明”：纯种犬[5]。犬种的标准化——主要以“地方品种”著称——成为民族自豪感的新来源[6,7]。到20世纪初，城市（伴侣）养犬已成为一种流行的爱好[8]。从乡村工作犬类型（牲畜护卫犬、牧羊犬和多用途农场犬）以及贵族猎犬——这些在一个世纪前大多不为公众所知[9]——中，第一次世界大战的剧变催生了一类新型实用/服务犬[10]。这些“现代工作犬”迅速在公众和各类武装部队中广受欢迎，为犬作为人类工作和运动伴侣的更广泛利用开辟了新途径（例如[11,12]）。
自20世纪90年代以来，研究犬的行为日益流行[13]。行为学家的重要突破来自这样一个认识：犬的自然环境是人为生态位[14,15]，因为这一概念将研究者的焦点引向犬的那些社会认知和行为能力，这些能力使犬能够作为成功的共栖者[16,17]和伴侣[18,19]适应人类社会领域。对犬与人互动相关这一极其复杂的科学研究领域进行系统性总结超出了本文的范围。主要方向之一是基�于分子遗传学的犬进化起源研究[20,21]，其中比较行为研究——包括最亲近的野生亲属灰狼（Canis lupus）——可能揭示犬与人类相关的特征性社会认知特征的可能起源[22,23,24]。相当一部分实证研究聚焦于人工选择（血统：[25]；功能选择：[26]；工作品系：[27]）和环境效应（教养：[28]；训练：[29]）对不同犬群行为变异的作用。一个快速发展的研究领域将犬作为有前景的、具有生物学相关性的模型物种，用于研究人类多种疾病在遗传学（例如[30]）、生理学（例如[31]）和行为（例如[32]）层面的问题。最后，犬与人关系的特征和新出现的问题也为应用行为学研究提供了充足的机会（例如[33,34]）。犬的问题行为具有非常复杂和多层面的背景，其中环境（例如社会化：[35]；训练：[36]；创伤：[37]）和犬的遗传背景（例如[38]）均可影响行为表型。在本文中，我们将重点关注那些源于原本健全的品种典型行为/气质、犬主期望与环境条件之间不匹配的潜在风险因素（例如[39,40]）。
品种标准和繁育组织仍然自豪地提及犬种的原始功能（例如[41]），但新时代的城市犬主通常并不依赖各类工作犬经过无数代选择所具备的那些才能和功能特异性行为特征。尽管纯种犬繁育的严格隔离原则有效地保留了许多功能性工作犬选择的主要行为特征（例如独立型与合作型工作犬品种之间的差异：[42,43]），但许多犬种可检测地分化为“展示”和“工作”品系（例如[44,45]），标志着与传统的功利性犬繁育方法相比，定向选择的方向发生了转变。
可以肯定地说，迈向21世纪，伴侣犬的饲养已成为一种生活方式[46]。据估计，全世界有超过两亿只狗作为宠物饲养[47]。此外，基于多项问卷调查，现代养狗者越来越将他们的犬类伴侣视为家庭成员，甚至视为“孩子”，而非仅仅称它们为家畜（例如[48,49]）。在工业化国家，养狗的普及程度已达到如此水平，以至于一些科学家甚至开始推测，伴侣犬可能充当其主人的替代品，而非主人自己生育孩子[50]。近期研究对此提供了更详尽的细节，表明伴侣犬为女性提供了比孩子更多的社会灵活性和与社交环境的联系；然而，母亲报告的对孩子的责任感高于对狗的责任感[51]。因此，尽管人与狗之间的情感纽带可以像人与人之间的情感关系一样复杂而深厚，但伴侣犬的社会地位也受到其主人家庭结构的影响。
城市化是一个全球性进程，其中人类居住密度的增长比例不断加速，为狗与人的共存带来了新挑战（例如[46]）。如在“西方文明”中，大多数饲养的狗被作为伴侣饲养（例如[52,53]），城市居民要求养狗的权利，即使这些动物的传统工作角色在大都市中无法实现。最近，城市规划和人类环境研究已开始评估城市养狗的条件、可能性和局限性，作为众多需要考虑的问题之一（例如[54,55]）。虽然如果狗至少部分时间在户外饲养，其对活动和环境丰富的自然需求更容易满足（例如[56]），但城市养狗给狗主人带来了严峻挑战。毫不奇怪，城市规划者、立法者和动物福利专家密切关注城市养狗“最佳实践”的许多方面，从每日最低散步次数到最长独处时间[57]，如何处理狗产生的废物[58]，以及 nuisance barking（扰民吠叫）问题[59]。
由于狗的行为可能是人与狗之间相互愉快共存的主要同步因素[19]，狗行为的许多方面因其对城市养狗的适宜性而受到显著关注。重要的是要认识到，各种工作犬品种中典型甚至偏好的特征，在今天可能被视为城市伴侣犬所不需要的。尽管关于什么是“理想狗”的细微细节可能因具有不同文化、经济和历史遗产的国家而异[60]，但高活动水平、保护与守卫、作为警觉且响亮的看门狗、以及倾向于追逐和狩猎，只是城市伴侣犬中很少受欢迎的特征的几个例子。
近期（城市）养狗的激增明显导致了“最终用户”（狗主人）以及社会对狗行为特征的新需求[61]。在整个20世纪，关于某一犬种对特定主人的适宜性的建议仍局限于将犬种的体型、性情以及可能的用途与主人的需求和生活方式相匹配（例如[62,63]），而最近，更具体的期望已被勾勒出来，暗示需要一种全新类型的狗。澳大利亚狗主人关于他们在“理想伴侣犬”中偏好的特征的反复调查意见[64,65]， crystallized（明确了）“友善”犬类型的形象。这可以描述为一种低驱动力、健康、对儿童安全、不吠叫且无攻击性的动物，无论品种如何，都应是城市式养狗的完美伴侣。尽管一些亚文化仍然偏好典型的不友善犬，并使用具有攻击性、威胁外观的狗作为地位象征[66]，但全球主导趋势倾向于友好和安全的城市伴侣。基于此，可以假设场景最终为犬种定向选择的新时代奠定了基础，以满足最大潜在狗主人群体的需求。本文的目标是批判性地评估这一矛盾情况的潜在解决方案，其中纯种犬的传统保护主义世界面临着自一个半世纪前起源以来可能最大的挑战。
2. 方法：文献调查
我们并非旨在撰写一篇系统综述；然而，为更好地充实本观点文章的背景，我们于2025年5月20日至23日期间进行了一次文献检索。我们使用四组关键词探索了Google Scholar和ScienceDirect：“dog breeds AND future”；“expectations AND dog breeds”；“change AND dog breeds”；以及“improving dog breeds”。随后于2026年8月14日至16日期间，在Google Scholar和ScienceDirect上进行了补充文献检索，使用了两个新关键词组：“urbanisation AND dog keeping AND demands”；以及“urban dog keeping AND problematic behavio(u)rs”。每次检索，我们均审查了前100条命中结果的标题和摘要。基于我们以往文献调查的经验，每次检索纳入更多命中结果主要导致重复条目和主题不适宜的条目。我们排除了非同行评审条目、学位论文和会议摘要，并保留了专著、书籍章节和同行评审期刊论文以供进一步评估。我们的主要兴趣在于犬行为、品种历史及犬与人互动，并额外关注与城市养犬相关的具体问题。在排除所有重复条目和范围外条目后，最终保留的论文集合共包含N = 60篇出版物。我们通过审查初步检索中部分最契合文章的引用参考文献，进一步丰富了文献收集。
3. 犬的类型、犬品种及相关行为特征：基于血统与功能选择的简要介绍
根据我们目前的理解，犬品种基于三个独立维度来定义和相互区分：(i) 其血统（即纯种犬必须是已知且隔离的奠基者群体的后代）；(ii) 其功能（即原始目的驱动的行为）；以及(iii) 其类型/外观[41,67]。重要的是要理解，这三个组成部分仅在维多利亚时代才成为纯种犬繁育不可或缺的部分。封闭血统谱系和由品种标准定义的犬外观严格参数在19世纪之前并不存在，各种工作犬类型（地方品种）的形成主要由基于行为的功能选择驱动（例如[3]）。自那时起，情况已完全改变。尽管犬品种据称的原始功能在主要犬业俱乐部组织的命名法中仍被强调[41]，但现代犬繁育者对功能特异性行为的重视程度降低，导致主要为 conformation shows [68] 和犬类运动而繁育的各种品种中出现日益均一、缺乏特征的行为性状。相应地，取代了原始针对理想功能特异性行为的强定向选择，对犬品种（往往极端的）视觉外观的更为相关的选择开始成为过去一个世纪犬爱好者世界的特征[69,70]。
尽管导致众多工作任务的特定犬类型（并最终形成品种）的传统功能选择压力在当今大多已放松，但很容易看出犬品种的主要行为/气质特征仍可表现出显著差异。最近，一项大规模分子遗传学研究未能检测到品种特异性行为表型[71]，但同一篇论文强调了几个跨越多个犬品种的重要遗传决定特征（例如“服从性”），将它们沿该行为性状的强度量表排列。这些发现表明有必要将焦点从品种水平的行为比较（例如[72,73,74]）转向寻找和研究能够影响更大品种集群的社会认知和行为表型的更广泛解释因素[41,75]。
当我们想要理解为何特定行为表型会在某些犬群中出现时，应始终关注生态效度（即特定表型在特定情境下为何会对研究对象有利），因此功能性品种选择（如[43,76]）为检验犬品种集群间生物学相关差异提供了一种特别合适的方法。我们应注意，犬的行为，尤其是其与人类及其他生物的社会互动，在现有犬的评估[40,77]和未来犬的“工程化”培育（如[64,78]）中具有决定性重要性。塑造各种犬品种类型的人工选择努力可能潜在影响工作犬的基础社会认知能力，这取决于其预期用途。“独立型”与“合作型”工作犬[79]之间的比较基于这些犬与主人之间不同的互动风格。合作型品种最初在密切的人类指导下以定期指令执行其工作（如猎枪犬、牧羊犬）。另一方面，猎犬、牲畜护卫犬（LGDs）和梗犬被选择用于必须独立执行的任务，要么在驯导员视线之外，要么无需定期指令。从为现代“最终用户”（即城市伴侣犬主人）寻找合适犬的角度来看，犬品种的功能性聚类具有特定优势。合作型和独立型犬类包含多种品种，形态特征高度可变。由于外貌是选择伴侣犬时最重要的偏好因素之一（如[80]），合作型和独立型犬品种类型内具有可预测行为特征的不同品种组合对潜在养犬者具有相当大的价值。犬的可训练性[81,82]和互动性[83,84]是伴侣犬主人的重要资产。与这些特征相关，研究发现功能性品种选择影响了犬的社会认知能力，这些能力也可能在日常犬与人互动中发挥作用。合作型犬品种更容易与人类建立眼神接触[85]，它们更成功地跟随人类提供的视觉指向[79]，并且它们也通过观察人类示范者学习得更好[42]。与此同时，研究发现独立型品种在面对困难任务时目光交替次数更少[86]，在主人看不到时更可能偷取禁食零食[87]，并且总体而言，它们对人类言语交流的关注程度低于合作型品种[43]。
针对犬类各种工作任务的人工选择从未停止；它可能只是近来转向了不同方向，例如：选择犬类用于新功能，如引导视觉障碍人士[88]；作为辅助犬[89]；或探测爆炸物的气味[90,91]。然而，各种犬品种类型之间的分化并非仅仅由人类的育种努力所驱动。基于分子遗传学，当今的犬品种（以及自由放养和野化犬）[92,93]可以根据其与共同祖先物种的遗传距离进行有效聚类[94,95]。除了基于功能选择的犬品种聚类外，品种支系与特定行为特征之间的祖先相关关联代表了行为研究的另一种主要的、具有生物学相关性的方法[41]。关于本文的主要议题，我们应当看到，从犬品种作为伴侣动物的适宜性角度来看，与所有犬假定祖先的遗传距离与一些最重要的社会认知和行为特征高度吻合。例如，古老品种（如Chow Chow、Akita、各种雪橇犬品种等）表现出比更多衍生（“现代”）品种支系更低的训练性[96]和更高的陌生人导向恐惧[97]。古老品种还表现出社交性与反应性之间更为一致的关系[23]。据作者所述，这一结果可以用驯化综合征来解释（如[98]），因为这些品种在很长时间内保持相对未改变，它们表现出更为一致的驯化效应。在更多衍生的品种中，这两个特征更加分离，这突显了在现代品种中[94]，社交性和反应性在针对特殊工作需求的选择过程中具有不同的重要性。除了个体行为特征外，品种群体更复杂的问题解决能力可以基于其祖先进行比较。通过观察人类示范者，现代品种，尤其是实用犬，在社会学习方面表现出比更古老的品种更高的表现，后者可能未经过强烈的定向选择以表现出与人类的高度互动性[25]。显然，当人们意图就犬品种基于祖先的行为特征得出有效且高度预测性的结论时，品种的广泛代表性和更多远缘品种支系之间的有效比较仍然是至关重要的必要条件。例如，当在一项调查中科学家仅研究“古老”犬品种的训练性时，其所谓的低表现若没有与更多衍生犬品种进行比较的可能性，几乎无法解释[99]。
4. 适宜与不适宜的行为：偏好行为与问题行为及其与品种/类型的关联
如果我们考虑到主要犬学组织认可的犬品种超过400个，那么决定为下一个犬类伴侣选择哪个品种并非易事。根据调查，每个品种或品种类型的可预测行为是潜在主人选择犬时需要考虑的最重要特征之一（如[80]）。尽管犬品种描述（“标准”）主要强调品种外貌特征，但它们通常也列出品种的原始（工作相关）功能和典型行为。然而，伴侣犬主人、兽医和犬业俱乐部在品种行为的有利和次优特征及方面可能有着非常不同的优先考量[11,44]。
例如，在澳大利亚社会中，理想的犬可以被描述为对儿童安全、易于室内训练、健康、服从且中等体型[64]。该研究发现养犬者与非养犬者的偏好存在差异，性别之间也存在差异：女性偏好更安静、无攻击性的犬，而男性则选择忠诚且精力充沛的伴侣。这些偏好在时间上相对稳定（参见澳大利亚的重复研究[65]）。另一方面，各品种受欢迎程度的强烈波动既未显示与消费者的犬行为偏好[100]相关，也未显示与健康、寿命、攻击性、训练性和恐惧性等品种特征[101]相关。品种受欢迎程度的趋势可能在不同历史时期受到不同类型信息的影响（如犬作为电影明星[8]、出现在电视广告中[102]、或与名人相关联[103]等）。
犬品种的培育主要是为了执行某种功能[104]。虽然这对于工作犬（如牧羊犬）和现代实用犬种（如导盲犬、警犬或嗅探犬等）来说似乎显而易见，但即便是玩具犬种也是基于功能方面进行选择的：为日常生活提供体型小、便于携带的伴侣犬种[105]。从这个角度来看，可以假设某一品种或相似品种群体所特有的行为可能对其原始功能具有优势。因此，在过去，人类很可能对这些曾经适宜的特征施加了强烈的定向选择压力[5]。牲畜护卫犬需要具备警觉、独立和攻击性的天性来保护畜群免受盗贼和捕食者的侵害[106]；梗犬的高活动水平和勇敢性被偏好用于消灭在鸡舍周围觅食的害虫[107]；而雪橇犬只要在雪橇前表现出色，就可以独立且不知疲倦[108]。
然而，这些因不同类型工作而被偏好的特定特征并不一定符合城市伴侣犬的生活方式需求（例如，对其他宠物、野生动物甚至小孩构成威胁的狩猎行为：[109]）。一旦养犬成为一项休闲活动而不再需要与工作相关的功能用途，问题行为便开始频繁出现，这主要是因为与原始工作功能相关的行为和气质特征并不总是适合作为伴侣犬。这一过程通过英国伦敦数百年来的记录得到了充分印证，显示了曾经有用的看门犬如何变成了如今的扰民吠叫犬[54,110]。当然，也可能存在某些品种类型更容易成为以友善行为为主的伴侣犬（例如，基于对澳大利亚犬主复杂偏好的评估：[65]）。然而，也可能存在某些品种类型远不适合成为家庭环境中的伴侣犬。问题行为通常具有复杂的病因背景（例如，分离相关问题，[111]）。因此，试图将这些因素彼此区分开来通常需要大样本的犬只。例如，研究发现问题行为背后的一个主要分组因素是犬的体型[112]。作者发现，除训练性低之外，每一种问题行为在犬品种体型越小的情况下发生频率越高。然而，犬的体型可能与其他因素如功能、饲养条件和寿命等相混淆[113]。因此，这些发现中的许多仍需要进一步研究。
在问题行为中，对人类缺乏攻击性是犬主在选择伴侣犬时考虑的最重要特征之一[82]。例如，梗犬，尤其是Staffordshire Bull Terrier，在公众看来甚至比立法者的评价更为危险[114]。社会对某些犬品种的这些往往是轶事性的认知需要通过实验证据进一步证实。在一项问卷调查研究中，Asp及其同事发现，来自AKC工作犬组的犬品种比非工作犬品种表现出更多的攻击行为[75]。另一方面，基于实证研究，无论品种是否受立法管制，品种内攻击性和反应性可能存在很大的个体差异[115]，这是人们可能反对禁止某些犬品种的常见原因[116]。
在受偏好和不受欢迎的特征背后，可能存在额外的混杂因素。研究发现，更恐惧的犬也更具攻击性[75]，这可以解释该特征的部分个体和品种相关变异。在另一项问卷调查研究中，发现17个品种的噪声敏感性水平存在显著差异[117]，然而，犬的年龄、性别和对分离相关问题的敏感性等混杂因素也被发现具有影响。应仔细研究这些特征的遗传力，以确定哪些是最不具攻击性或恐惧性的个体用于进一步繁育。然而，由于大多数行为特征的复杂遗传背景，这可能是一项困难而漫长的任务。即使研究人员在一个品种中发现了遗传关联（例如，Havanese比熊犬的恐惧行为，[118]），这些发现在应用于不同犬品种时也可能无法推广[118]。
在考虑犬只可领养性时，攻击倾向是一个至关重要的特征。在做出领养决定前与一只行为不良的犬短暂接触，可能会不公正地给同品种犬只带来不利影响[119]。对领养犬只不满意的主人可能会将犬只退回收容所（即使没有问题行为的具体证据[120]）。反过来，纯种犬价格更高，主人宁愿忍受不便[121,122]，无论是否尝试通过训练改善共处质量。对于混种犬，领养前由犬类专家提供咨询可能是提高领养成功率的好办法[123]，因为研究发现，放弃犬只的主人对应有行为和犬与人之间纽带的期望更高[40]。此外，对收容所中犬只进行行为评估可以大大提高犬只的可领养性[124]，并可能减少对人的安全隐患。
问题行为也可能由于在 conformation show 和宠物犬繁育过程中缺乏严格的行为选择而发生。而在工作犬繁育中，人们高度重视提高选择效率[27]，对于缺乏工作功能的犬种则不能这样说[125]。尽管繁育者对犬种的兴趣日益增加且观念发生了转变[126]，但特定犬种可以预期什么样的行为和气质已变得越来越不可预测。这一趋势得到了一些繁育者态度的支持，他们——尤其是在高度流行的犬种中——继续饲养和繁育已有充分记录的慢性健康和行为问题的动物（例如，牛头梗的“trancing”行为：[127]；短头犬的呼吸问题：[128]）。某些犬种的高近交水平或遗传变异性低的犬种，也不一定有利于行为稳定性[129]。然而，高水平的近交在被认为是“健康”和“不健康”的犬种中都可以发现[130]，因此近交对行为和健康的确切后果应通过有针对性的研究来厘清。例如，研究发现近交可以影响犬与人互动相关的社会认知特征，如指示犬种中对人的厌恶[131]。
最后，大量据报道的问题犬行为可能是主人对偏好犬行为的期望不切实际的结果。许多人想要一个以人为中心、高度依恋的犬类伴侣[64]，然而，当他们白天离开犬只时，他们也期望这只高度依赖人的犬不应表现出分离相关行为[132]，并且在他们回家时不应过度吠叫。另一方面，偏好健康犬只的主人继续选择具有已知遗传性疾病和解剖学问题的犬种[82,133]。如何纠正和解决这些不切实际的期望，需要采取举措帮助主人认识到正确选择犬种的重要性（例如，[134]）。
5. 影响/改善犬行为和主人-犬匹配成功率的多层次尝试：品系选择、犬种杂交、幼犬测试
尽管维多利亚时代纯种犬繁育的标准化将重点转向了犬种的外观[1]，但犬的行为从未变得次要[75]。这不仅反映在继续将工作犬用于各种传统（例如，牲畜护卫[135]和放牧[136]）和新发明的工作（例如，各种辅助犬[89]和服务犬[137]）中，更重要的是，反映在作为愉快动物伴侣的渐进且最广泛的作用中，这需要持续维持行为健全性[78]。
当我们指出犬的自然栖息地是人类生态位[18]时，这也意味着犬必须适应人类环境中新出现的挑战[19]。当然，特别是在纯种犬的情况下，适应主要通过人工选择发生（例如，[138]），但与自然选择下的动物种群相比，结果基本上是相同的：犬种的形态和行为特征正根据对人类提出要求的更优适应而被塑造[41]。无论我们是考虑为特定目的需要合适的工作犬[27]，还是考虑作为兼容宠物的角色，这仍然是正确的。伴侣犬主人根据其生活方式[139]、时尚[102]和环境限制[80]有明确的需求，但犬的行为可能始终是最重要的表型，最终决定犬与人之间的匹配是成功还是失败（例如，[40,123]）。毫不奇怪，人们发明并继续使用大量方法，目的是改善或确保犬只理想的行为特征，使其适合特定目的。这些程序范围从对期望行为特征（或反对不期望特征）的定向选择，到不同犬种的计划杂交，再到通过早期评估和短期/长期训练方案针对个体犬只行为的干预。
5.1. 我是留下还是……消失——有利犬行为定向选择的潜力
在品种标准中，犬通常以其被认为“符合类型”且根据其原始用途而受偏好的行为特征来描述。颇为有趣的是，在所有引起科学兴趣的犬类行为中，那些不受欢迎的、甚至“异常”的表型是最早被认为具有遗传背景的行为之一。其中，我们可以提到指示犬的“不稳定性情”[140]，或红色/金色英国可卡犬中的无端攻击行为（“可卡狂怒”）（例如[141]）。自那时起，除少数例外[64,78]，伴侣犬遗传行为的研究背景仍然以对各种不良表型的研究为主[142]，而应用动物行为学家则更可能关注实用/工作犬的偏好行为特征[143]。
在一项基于瑞典BPH测试（犬行为与性格描述）数千只犬结果数据库的研究中，Tygesen等人[144]发现了五个具有弱至中等遗传背景的行为因子：（1）嬉戏行为；（2）逃跑与保持距离行为；（3）社交性；（4）社交威胁与恐惧行为；（5）顺从行为。利用八个样本量充足的品种（美国斯塔福德郡梗、金毛寻回犬、拉布拉多寻回犬、拉戈托罗马涅罗犬、新斯科舍诱鸭寻回犬、罗得西亚脊背犬、西班牙水犬和斯塔福德郡斗牛梗）的数据，他们发现逃跑与保持距离行为具有最强的遗传力，并认为其结果可能对未来培育更少恐惧、更易训练的犬只有用。
各种形式的攻击行为是研究最为深入的不良犬类行为之一，其中，除环境影响外[38]，我们可以找到大量关于遗传易感性的实证证据（例如[145,146,147]）。焦虑相关的行为问题在伴侣犬群体中也很普遍[148]，因此，研究人员试图寻找可能促成这一特征的遗传因素也就不足为奇了（例如[149,150]）。
大规模分子遗传学技术（如观察性和潜在性状GWAS）的可及性和成本效益不断提高，使科学家能够分析大量样本的遗传物质。Mahmoodi等人[151]研究了来自100多个品种、超过4万只犬的样本，寻找源自广泛使用的C-BARQ工具（例如[104]）的九个行为特征的候选基因。这些都是在伴侣犬中被认为“有问题”的行为：对陌生人的攻击行为、对主人的攻击行为、对犬的攻击行为、对陌生人的恐惧、非社交性恐惧、对犬的恐惧、触摸敏感性、分离相关行为和依恋寻求关注行为。结果显示，在识别出的41个显著SNP中，有26个属于多效性神经行为基因，这凸显了针对所列表型进行潜在行为选择工作的复杂性。
尽管迄今为止，伴侣犬遗传影响行为特征的研究主要集中于有问题的表型，但这些因素的识别显然可以帮助繁育者选择表现出更有利行为模式的犬只。在犬品种内朝着“理想”伴侣犬方向进行定向选择将是一项值得称道的努力（正如一些专家所建议的，例如[78,125]），特别是因为根据那些近期将“工作”线和“展示/宠物”线分开的品种的结果，工作型犬通常比非工作型犬表现出更有利的行为维度。D'Aniello等人[29]发现工作型拉布拉多寻回犬对人类更加关注，Fadel等人[44]发现工作型拉布拉多犬和边境牧羊犬比同两个品种的“宠物线”犬保留了更多品种典型的冲动性特征。一种解决方案是确定并追求具有最有利“伴侣犬特征”的繁育目标（例如[64,65]）。然而，理论上还有另一种方法可以提供更快速（但仍基于遗传学）的配方来达到所追求的、理想的伴侣犬行为：通过品种间杂交来组合现有的行为表型。
5.2. 为“良好行为”而设计犬？品种杂交
尽管从定义上讲，纯种犬繁育依赖于根据犬种登记册所确定的“纯血”血统，但在过去150年里，一些新品种犬已从现有品种或地方品系中培育出来。或许通过精心选育和杂交当代地方品系而创造出的最著名品种例子就是德国牧羊犬[152,153]。金毛寻回犬是另一个在爱好者和专业人士中获得了持久欢迎的品种，其起源可以追溯到19世纪末对当时现有犬种的有意杂交，如纽芬兰犬、爱尔兰雪达犬、猎鹬犬和寻血猎犬[154]。在大多数情况下，繁育者对这些预先计划的杂交有明确的功能目标，他们的意图是获得一个“更好”的工作犬品种，结合原始（亲本）品种的有益特征。例如，猎兔犬[155]是由牧羊犬和视觉猎犬（如边境牧羊犬和灵缇）杂交产生的；由此产生的猎犬在理论上比更独立的视觉猎犬更能接受指令，因为它有牧羊犬祖先[156,157]。
由于每个现有犬种都可以通过品种典型行为的组合来表征（例如[41,75,96]），如果犬类繁育者将创造理想伴侣动物的努力建立在结合不同品种的有利特征上，那也是合乎逻辑的。一个关于杂交潜在效果的间接（且不推荐）例子来自犬狼杂交种的比较行为学。Tagliarini和Temrin[158]将犬狼杂交种对人的导向行为与阿拉斯加雪橇犬、西伯利亚哈士奇和德国牧羊犬进行了比较。每个研究对象在美国都作为伴侣动物饲养，根据DNA检测，杂交种具有11-43%的“狼含量”。虽然犬狼杂交种与雪橇犬相比表现出相似的玩耍性，且并不比研究中的犬更具攻击性，但它们比任何其他受试犬都显著更恐惧。这些结果表明，杂交种的行为在某些情况下可以根据我们对亲本形式的了解很好地预测：狼是未被驯化的野生动物，有着被集中猎杀的长久历史；因此它们对人类的恐惧[159]在其杂交种中同样占主导地位并不令人惊讶。有趣的是，恐惧/温顺是Scott和Fuller[160]开展的开创性项目中巴仙吉犬-可卡犬杂交中明显分离的行为表型之一，这表明该性状在犬中具有相当简单的遗传背景。然而，这些作者发现，杂交犬的其他能力，如解决问题的能力，更难以与亲本品种的行为联系起来。
到20世纪末，为创造受欢迎的伴侣动物而有意识地对已建立犬种进行杂交变得越来越流行。通常被称为“设计师犬”，它们通常是为了外观和其他形态特征而非行为而培育的[70]。最著名的设计师杂交犬集群是所谓的“涂鸦犬”，它们是贵宾犬与其他犬种杂交产生的杂交种（例如，拉布拉多涂鸦犬是贵宾犬-拉布拉多寻回犬的杂交）。贵宾犬杂交最初是为了制造低过敏性被毛[161]，但除了创造各种设计师犬的繁育者明显的金钱利益之外，据称积极的行为特征在品种杂交中也经常被强调[70]。
设计师品种的饲养者最常提到这些犬的外观是其选择背后的驱动力[161]。除此之外，预期的正向杂种优势（或杂交活力）效应——即相比传统纯种犬通常更健康的犬——也常被强调为设计师犬日益受欢迎的原因。据我们所知，这一假设尚未得到实证支持[162,163]；相反，品种杂交似乎表现出其亲本品种典型疾病的组合谱系[164]。但它们的行为又如何呢？除了极少数研究（如[138]）外，设计师杂交犬的行为尚未被评估。与几个世纪前的犬类繁育者或有目的地针对功能（工作）性能的现代行为遗传学家[143]不同，现代设计师犬繁育者显然没有专门旨在改变这些杂交犬相对于其来源品种所表现出的行为。基于孟德尔遗传学，可以假设F1品种杂交犬将表现出亲本品种的“中间型”行为表型。遗憾的是，迄今为止尚未有类似Scott和Fuller所做的经典项目[160]的此类研究，而Shouldice等人[165]的调查显示，Goldendoodles、Labradoodles及其亲本品种（拉布拉多或金毛寻回犬和贵宾犬）的C-BARQ行为特征[166]之间没有相关差异。从“理想伴侣犬行为”的角度来看，经过深思熟虑规划创造的设计师杂交犬可能具有巨大的潜在价值；然而，更深入地理解那些在犬与人互动中起关键作用（无论正面或负面）的行为的遗传机制将至关重要。
5.3. 窥探行为个体发育的水晶球：幼犬测试
犬一生的大部分时间都是作为成年动物与主人一起度过的；因此，犬与人类期望/需求之间的匹配主要基于成年犬的行为来评估。自Scott和Fuller[160]的开创性研究以来，获取幼犬的推荐时间窗口一直被认为在生命第八周左右。然而，由于近期有观点认为幼犬可以从其母亲和同窝兄弟姐妹的社会环境中大为受益[28,167]，如今有一种趋势是将幼犬交给新主人的时间不早于10–12周龄（如[65]）。同时，无论犬的未来角色如何（即工作犬或伴侣犬），以下情况将非常有益：(1) 基于幼犬的行为特征，能够更好地保证未来主人需求与犬之间的“完美匹配”；(2) 幼犬的养育能够预防日后出现的问题行为和普遍不健全的行为特征。从这个角度出发，专业人士和研究人员已经提出并测试了几种方法，用于改善幼犬仍与母亲在一起时的刺激环境[168]，或评估幼犬的行为以预测其成年后对特定目的/需求的适用性（如[169]）。
幼犬测试至少可以通过两种方式帮助预测后续（成年）行为特征。看似更简单的期望是，基于幼犬的气质/行为，可以将“合适的犬分配给合适的主人”（如[170]）。这些需求通常不会比试图预测某些幼犬是否会成为其品种中高度活跃/高驱动力的个体，还是较为悠闲的个体更深入地涉及犬的行为。根据研究，在这种情况下，缺乏经验 prospective 主人只需带着母犬一起参观幼犬窝，以了解该品种和繁育品系的预期行为，就会大受裨益（如[39,171]）。然而，有时通过对幼犬进行行为测试来预见更具体的信息将是可取的，例如它们是否会适合武装部队所需的某些任务，或成为高级辅助犬训练的成功候选者[172,173,174]。Junttila及其同事[175]发现，当他们对3至7月龄的幼犬进行测试时，其测试组合中的某些任务能够适当地预测成年犬的相应行为特征。跟随人类指向手势预测了成年期的可训练性，而在绕行任务中表现良好预测了能量水平和冲动性，“不可解任务”中的表现与成年期对陌生人的恐惧相关。尽管在该研究中，年轻受试者已经过了大多数犬进入新主人家庭的典型“幼犬年龄”，但通过适当选择的气质和个性测试，人们可以找到从幼年到成年相对稳定的行为特征（如[176]），这似乎是合理的。遗憾的是，在现实中，这些期望往往无法被证实（如[177,178]）。
这些令人失望的结果的一个原因可能是，幼犬的行为库和典型反应可能与我们希望预测的成年工作犬/运动犬的典型任务相关行为存在巨大差异（例如，[172]）。由于幼犬通常只在其出生的场所停留相对较短的时间，且繁育者和未来主人都希望避免将日后不适合期望角色的幼犬错误安置，因此“理想”的幼犬测试应尽可能早地实施[170,177]。这增加了此类程序的难度，因为犬只年龄越小，可测试的行为反应就越简单，而这些反应不一定类似于需要预测的成年行为模式[172]。因此，毫不意外的是，据我们所知，目前备受追捧的幼犬测试“圣杯”（“早期——简单——高度预测性”）尚未被发现和发表。
6. 未来的犬品种：对其发展的四种潜在途径的提议与评估
在前几章中，我们概述了现代（后维多利亚时代）犬繁育时代不仅创造了令人眼花缭乱的犬品种，而且在20世纪末还促成了世界各地伴侣犬饲养前所未有的流行[102,179]。城市犬饲养的兴起导致了一个悖论性的局面：人们现在可以从数百个已建立的犬品种中进行选择，这些品种具有足够广泛的形态和行为变异性，似乎能满足每个人的需求——然而，现代伴侣犬饲养仍受到行为问题犬只发生率快速增长的困扰（例如，[180,181]）。主人的期望、能力与所购犬只的气质/行为之间的不匹配，很可能导致主人的不满和负面体验、弃养动物，进而损害人与动物的福利[181]。我们已经看到，理论上手头有几种解决方案，这些方案可以帮助潜在犬主人做出最佳选择，并了解可以从他们的新犬伴侣身上期望什么。科学研究已根据犬品种的血统和功能选择突出了其若干行为和认知特征（综述见[41]），这可以在规划阶段就这些品种的行为提供基本指导。基于亲本品种行为的品种杂交可能具有一定的预规划潜力。对于同一犬品种中分离的工作系和宠物/展示/运动系，也有一些科学证据可用于对其预期行为做出有根据的推测[44,182]。我们还应该提到广泛训练方法不可或缺的作用和重要性，其中许多已被发现能有效改善友好的人犬互动的便利性和质量（例如，[168,183,184]）。然而，这些方法和信息来源已经存在相当长的时间，从接近“理想犬”的角度来看，人犬关系的质量并未取得任何显著进步。
有关纯种犬因极端解剖学特征而导致的健康问题及福利受损的争议与问题，已促成一些开创性出版物的问世，为未来的育种工作提供了潜在解决方案（如[70,125]）。理论上，多项近期调查也已汇总出一只高度成功的城市伴侣犬所应具备的基本特征[64,65,185]。据此，犬主更倾向于选择一只“友善”的犬：与儿童相处安全，不对人类和其他动物表现出攻击性，可长时间独自留在家中而无问题，不过度吠叫，且性情温和放松。尽管某些犬种的行为特征可能比其他犬种更接近这一理想化的犬类伴侣，但我们尚未了解到有任何科学研究或应用项目（工作犬项目除外，如[91]）曾尝试培育更接近“改良版”伴侣犬的犬只。由杰出研究人员和犬类专业人士领导的“功能性犬协作组织”（https://functionalbreeding.org/about-us/，于2026年8月22日访问）等倡议，进一步凸显了公众对一项以科学为基础的大规模项目之需求——该项目旨在选育在健康与行为特征上同样健全的犬只。鉴于问题性（不匹配的）人犬关系的高发生率及其令人担忧的后果[181]，且基于我们当前改善这一状况的可能性明显不足，我们认为社会应迈向以科学为基础的功能性犬育种新时代。在本章中，我们概述了培育此类犬只（“品种”）的若干潜在途径，这些犬只既能同等满足各类终端用户（一般为伴侣犬主人）的需求，又能为犬只自身提供良好的福利和安全的生活。我们还将借助Van Belle等人[186]为应用行为学研究概述的SWOT（优势、劣势、机会、威胁）原则，对每种拟议策略进行评估。该分析工具为多方面评估潜在“竞争性”解决方案（如[41]）提供了直观的框架，从而有助于为未来研究和决策提出考量建议。
四种拟议策略的直观图示见图1。
[图：实现行为可预测且适合的犬种四种拟议育种策略的示意图]
6.1. 策略“趋同”：使大多数犬种向“友善城市犬”趋同
这一途径已由那些调查澳大利亚和意大利公众对犬类伴侣行为偏好的作者提出[64,65,185]。一旦目标设定，理论上育种者基本上可对任何犬种施加强烈的定向选择，以尽可能接近所需的行为表型。可以假设，有些犬种从一开始就更接近理想城市犬的某些选定指标。其他犬种，尤其是那些被选育/用于护卫、个人防护、与其他犬只/动物搏斗，或表现出高度合作性（如牧羊犬）或独立性工作（如梗犬、腊肠犬）驱动力的犬种，最初将无法满足许多要求。我们还应牢记，除行为评估结果外，关于某犬种是否适合担任友好家庭伴侣的看法，可能受到与品种相关的刻板印象以及带有正面或负面体验的偶然接触的强烈影响[119]。
大多数研究致力于识别在恐惧、攻击性、分离相关问题及过度活动水平等稳健表型上最高和最低的犬种（如[75,187]）。McGreevy等人[112]发现，总体而言，犬种体型越小，其表现出的行为问题迹象（包括恐惧和攻击性反应，或与主人分离的问题）越频繁、越强烈。一项波兰研究强调，虽然体型较大的犬更常被报告有攻击性问题，但体型较小的犬种据报告比其他犬更“多动”[188]。被选育为与驯导员密切合作工作的犬只，比独立工作的犬种表现出更强的分离问题倾向[132]。古老犬种（如阿拉斯加雪橇犬、西伯利亚哈士奇和秋田犬）也可能表现出对其他动物强烈的攻击性、过度活跃和令人不安的吠叫[189]。因此，总体而言，每个犬种或品种类型从理想伴侣犬的角度来看都可能表现出不良特征。根据“趋同”策略，为培育未来理想伴侣犬，基本上每个犬种都需要中等到强度的定向选择才能实现这一目标。
这一未来犬种拟议途径背后的主要“理念”是，它基本上将几乎全部责任置于犬类育种者身上，并间接地置于犬只自身。理论上，一旦培育出所需的“用户友好型”犬只，未来的犬主便可安全地获得它们，而无需担心该犬种先前预期的行为。
优势：如果选育大量犬种以成为“理想伴侣犬”，这将显著提升“使用者”（犬主及其家人）以及其社会环境（其他人和动物）的总体安全性。我们可以基于以下认识预期这一点：对人的攻击性和对动物的攻击性是犬最严重的问题行为之一，且往往带来悲惨后果（例如[190,191]）。从大量犬种中广泛获得“友善”犬只将对养犬体验产生积极影响，因为问题犬只在人类环境中造成相当大的情感和经济负担[181]。我们还可以预期，更无压力的犬与人共处以及分离相关问题等问题行为发生率的降低，也将提升犬的福利[192]。最后，该策略的优势在于，我们已经积累了关于犬种行为各个方面的丰富实证知识，这可以为所概述的计划提供坚实的起点。
劣势：理论上，每个犬种都可能具有遗传灵活性，允许朝着友善伴侣犬的要求进行定向选择，我们也不应忘记犬主额外训练作为进一步接近理想伴侣犬的积极效果（例如[33]）。同时，便利样本犬种之间的大规模行为比较（例如[193]），以及基于功能性犬种选择[194]或犬种群体祖先（例如[189]）的犬种类型比较表明，犬种的典型行为和社会认知能力之间存在足够强烈的差异，这可能使得从特定犬种或犬种类型中选择理想伴侣犬成为近乎徒劳的努力。尤其是今天，那些曾被认为对某一犬种功能最有用的不良特征似乎难以通过选择来“淡化”。例如，护卫犬种的高度警觉、吠叫和领地行为，或梗犬的“多动”，是原始工作导向选择的此类标志，在理想的城市伴侣犬中将是次优的。最有可能的是，只有一部分当今犬种可以为进一步选育理想城市伴侣犬提供容易的起点。由于主要目标性状（攻击性/恐惧性、多动和分离相关问题的易感性）可能与有些不同的遗传背景相关联（例如[147]），同时针对每一个性状进行选择可能非常困难。因此，育种者可能不得不选择其犬种的哪个特征应首先改进最为重要。随机繁育的混血（“杂种”）犬代表了“趋同”策略框架中的另一个弱点。混血犬在各地伴侣犬中占很大比例（例如，在美国，据报道近50%的犬群是非纯种犬[195]）。对于混血犬，我们没有关于犬只预期行为特征的广泛先验知识；因此，利用它们进行理想伴侣犬的有效定向选择可能是一项极其困难的任务。
机遇：“趋同”策略可能为那些因其原始工作功能而变得不受欢迎或近乎灭绝、从而其某些行为特征被证明与城市伴侣犬饲养不相容的犬种提供“最后的机会”。大型牲畜护卫犬种（LGDs）最近被重新引入以执行其原始任务，即保护畜群免受狼、郊狼和金豺的威胁（例如[196]），否则将难以融入伴侣犬身份。这些犬种由于其体型大和过去的选择而容易伤害人（例如[197,198]），部分因此，它们作为伴侣犬的受欢迎程度正在下降。“趋同”策略可以鼓励犬只育种者开发更适合伴侣用途的LGDs版本。同样的机遇也适用于其他具有强烈保护倾向的犬种（例如Rottweiler、Boxer、German Shepherd Dog、Dobermann等）——尽管这些犬种没有灭绝威胁，但由于其对人攻击性的报道（例如[199,200]），它们可能存在问题。“趋同”策略的机遇将得到特定犬种中已经进行的分裂选择过程的助力，其中存在不同的“工作”和“展示”品系[44]。这些品系可以揭示犬种内的行为变异性，并容易为进一步选育理想犬类伴侣提供更多原材料。从潜在犬主的角度来看，“趋同”策略还可以为那些偏好因其负面声誉而原本在社区中不受欢迎的犬种的人提供有利结果（例如[201]）。最后，由于“趋同”策略必然包括适当行为测试的开发和应用，这些测试也可用于混血犬，从而为收容所工作人员、救援组织和潜在犬主提供关于这些犬只作为城市伴侣犬适宜性的见解。
威胁：“趋同”策略间接对未能表现出向理想伴侣犬方向理想改良的犬种构成威胁。如果潜在饲养者普遍期望现有犬种具有可预测的“友善”行为，那些犬种将面临人气迅速下降，其中繁育者不愿选择适合家庭的品系，或缺乏所需遗传背景削弱了繁育者的努力。此类犬只在从收容所领养时也会面临困难（例如[202]）。“趋同”策略中一个仍未知程度的威胁是，由于向伴侣犬方向的选择，某些犬种典型行为特征可能不可预测地丧失。有迹象表明，一些犬种的“展示品系”在过去几十年缺乏原始工作相关行为选择，它们表现出的行为特征彼此之间比其工作品系对应者更为相似[44]。一旦这些功能特征丧失，我们可能面临由此产生的犬只也失去其先前吸引力的很大一部分的情况。在“趋同”策略广泛成功的情况下，原始（功能）工作犬的使用者可能面临难以找到足够大的其喜爱犬种此类犬只来源的困难。另一个不同性质的潜在威胁可能是与“友善性”某些受青睐特征有遗传、神经或生理联系的潜在不期望表型的很大程度上意外的出现。一个功能类比由仅以温顺性选择过的银狐研究提供；然而这种定向选择也影响了它们的形态和非目标行为，如它们的发声[203,204]。
6.2. “分化”策略：通过加强犬种典型特征来减少犬种行为不可预测性
犬的大多数行为表型受环境（即早期社会化[28]、训练[205]和既往创伤[206]等）和表观遗传学（例如[207]）的强烈影响。然而，近期分子遗传学进展（例如[208]）仍表明主要工作类型犬（如放牧、捕食和可训练性）的标志性行为特征存在可检测的遗传背景。“分化”策略背后的核心思想与当今潜在犬主普遍推荐的“最佳实践”并无不同：在仔细审视可能性和了解犬种现有特征后，才从最匹配的犬种中做出明智选择[80]。当前研究表明，犬主确实使用多种信息来源了解犬的特征和对主人的要求（例如在英国：[209]）。然而，正如Mead及其同事[209]所指出的，这些关于犬种的信息往往存在争议，而根据另一项研究，“一时兴起选择犬只”仍是主人最常报告的方法之一[82]。如果我们再加上犬种常常差异很大的行为特征（例如[210]），并因基于繁育品系的歧化选择导致的分化而进一步强化[44]，“分化”策略背后的理由就变得显而易见：如果犬种表现出更“清晰一致”的行为特征，这种可预测性可以积极回报那些原本密切关注使其需求与可能性同所选犬种特征相匹配的犬主。值得注意的是，“分化”策略使用与“趋同”策略不同的繁育技术（稳定化选择与定向选择），并且对主要责任的强调也不同。“趋同”策略将主要责任放在繁育者身上，而“分化”策略则突出潜在犬主的责任，并通过阐明各犬种和犬种类型之间行为特征的（可预测的）差异来为他们提供帮助。
优势：也许“分化”策略可以建立在关于各犬种行为特征的现有经验知识的最广泛基础之上。有丰富多样的研究和综述论文从分子遗传学（例如[71]）、功能选择[41]或基于遗传距离的祖先[211]方面探讨了与犬种相关的行为差异。基于对原始工作任务和犬种满足这些要求的行为特征的实践描述，借助应用行为学在行为评估中的工具包（例如[25,68,193]），繁育者可以选择行为表现最“适合该犬种”的个体进行进一步繁育。同样的“犬种刻板”知识可以为潜在犬主提供指导，他们将从应适合其期望的犬种中获得此类犬只。
弱点：“分化”策略的主要问题之一是，许多初次养狗者和有经验的狗主对负责任的养狗知识[212,213]以及特定犬只（品种）的需求和特征了解不足。研究表明，在选择下一只狗时，极端短头品种的犬主即使意识到这些动物健康和福利受损，仍会再次选择同一品种[133]。这类犬主愿意接受并将例如“懒惰”正常化为巴哥犬或斗牛犬的品种特异性特征，而非将其视为心脏和呼吸功能障碍的严重后果[214]，这种态度对于其他有问题但被普遍“接受”的品种相关行为和解剖学异常也可能是一种普遍现象。因此，“分化”策略的成功主要取决于对犬主进行有效教育，使其了解每个品种的理想和典型预期行为是什么。“分化”策略的另一个弱点是许多品种行为的广泛变异性——这种变异性因快速分化的“工作”和“展示”血系而进一步加剧。例如，Overall等人[215]发现，噪声敏感性不仅表现出品种差异，而且在澳大利亚牧羊犬、边境牧羊犬和德国牧羊犬中还表现出家系差异。品种内强烈的行为变异性要求在育种组织就“真实”品种特征达成一致时需极度谨慎。最后，“分化”策略还有一个弱点，即缺乏为混血犬建立理想典型行为特征的指导原则。
机遇：与“趋同”策略相比，理论上在此策略下每个品种都可以以其原始功能形态延续——只要社会和“最终用户”仍有需求。如果犬主能获得关于其需求和特征的可靠有用指导，并接受关于自身作为犬主的能力和责任的现实评估教育[212]，大多数品种很可能能够适应伴侣犬角色。“分化”策略成功的关键在于准犬主做出正确选择，并使这些选择影响育种者按照品种标准提供行为可靠且可预测的犬只的努力。如果遵循“分化”策略建议的育种者成功稳定了品种的特征性行为性状，不仅有利于准伴侣犬主做出明智选择，工作犬使用者也可从中受益[27,143]。尽管许多行为性状的遗传倾向（遗传力）仍知之甚少，且个体发育和环境因素的作用不应被低估[27]，但如果工作犬育种者和训练者能从行为特征更明确的品种入手，其任务可能会更容易。关于混血犬，由于“分化”策略未指定理想伴侣犬的普遍行为蓝图，混血犬广泛且不可预测的变异性在经过适当行为特征分析后也可加以利用。如果能有关于混血犬主要行为性状的更可靠信息，偏好领养混血犬的准犬主也能找到更匹配的伴侣。Archer等人[216]表明，在提供犬只照片的同时提供性情/行为特征描述，比提供附有推测品种标签的相同照片更能积极影响收容所混血犬的领养率。
威胁：“分化”策略一个潜在且令人担忧的结果是，如果育种者利用稳定化选择来选育行为性状而未采取有效保护犬只遗传变异性的措施（如[217,218]），品种内近交水平可能进一步升高。“分化”策略并非追求通用的友善行为特征，而是强调强化品种原有的、适合功能的行为。因此，那些原始用途已过时且典型行为主要不适合伴侣动物用途的品种，更可能消失。“分化”策略最相关的弱点可能是缺乏明确界定的“理想城市犬”，因为该策略依赖于受过教育且准备充分的准犬主（理想化）形象，即他们能从行为可预测且可靠的品种中做出正确选择。考虑到当前广受欢迎的犬种也普遍存在一长串基于形态学、遗传学和行为学的健康、福利和安全问题[101,102]这一令人失望的趋势，期望犬主选择最符合自身及其环境条件的品种可能是不切实际的。“分化”策略的另一个弱点是，如果对犬主和育种者的教育有误或效果不佳，不当强调品种特异性特征可能产生不良效果，即进一步推广/正常化某些品种的行为和形态极端特征。
6.3. 策略“拆分”：在同一品种内创建易于区分的工作型和城市伴侣型变体
如果我们考虑“趋同”和“分化”策略的主要威胁，就会清楚地看到，前者间接意味着原本受重视的工作犬行为特征最终会丧失，而后者可能无法提供广泛选择的理想城市兼容伴侣犬品种。“拆分”策略可以视为前述两种方法的结合，其双重目标是改善目标品种的行为。一方面，目标是培育出可靠的“友善”类型；另一方面，根据现有种畜的实际行为状况，恢复或维持可靠的工作品系。
对于“拆分”策略能够提供普遍有用解决方案的犬种，大型牲畜护卫犬（LGDs）是一个很好的模型。根据最近的分子遗传学结果，这些品种（最初为地方品种）在欧亚大陆各地被培育出来[219]，基本上是为了同一目的：保护各种牲畜免受危险捕食者的侵害和被盗[220]。这些犬表现出极端的忠诚，对牲畜没有捕食行为[135]，但可以理解的是，它们上述与工作相关的行为特征（以及大型体型）很难适合城市伴侣身份。当大型捕食者（如狼）在何时何地被猎杀至灭绝时，LGDs失去了原有的工作，其数量以及管理知识也大幅减少（例如[221]）。当由于实施的生物保护计划而使天然捕食者物种开始返回其原始栖息地时，对可靠LGDs的需求再次出现[196]。目前，LGDs不仅被认为是大规模牲畜管理中最有效，同时最经济且对野生动物最友好的解决方案之一[222]。然而，如今可能更多的LGDs是出于爱好（用于展示和/或作为伴侣）而繁殖，而非用于保护牲畜。大多数LGD品种被认为不太适合作为家庭伴侣（例如[223]），尽管它们令人印象深刻的外观和体型吸引了许多人[82]。理论上，对于像LGDs这样仍然需要履行原始角色、而该角色又与伴侣身份不完全兼容的犬种，“拆分”策略可以为两种利用类型提供令人满意的解决方案。
优势：“拆分”方法的主要优点是，它基本上可用于任何犬种，比“趋同”或“分化”策略具有更大的灵活性。需要进行初始的全种群评估以明确：（1）是否需要保留品种原有的功能特异性行为；（2）该功能是否与理想伴侣犬行为库的要求显著不同。采用“拆分”方法的犬种可能在哪个育种目标（“伴侣” vs. “原始工作”）需要更多育种努力才能达到方面存在差异。“拆分”策略的采用者可以依赖某些犬种中已有的育种品系（“展示”或“工作”品系，例如凯尔皮犬：[224]；边境牧羊犬和拉布拉多寻回犬：[44]）。
劣势：目前尚不清楚，在一些已经针对特定任务/性情进行了强烈选择的犬种中，是否还剩下足够的遗传变异，以通过朝着完全不同的功能角色育种来取得可观的结果。许多犬种遭受严重的遗传多样性丧失（例如兰伯格犬：[225]，杜宾犬：[226]），这表明在每个品种中为不止一种行为类型寻找可持续的广泛遗传基础可能是有问题的。该策略的另一个普遍劣势是，它不为混种犬在育种或评估计划中提供位置。
机遇：与“趋同”和“分化”方法相比，“拆分”策略可能会在各种犬种的爱好者中获得最广泛、最容易的接受。与将整个品种推向预定义的统一行为目标不同，采用“拆分”策略理论上可以建立同一品种的“伴侣”和“工作犬”版本。这不仅能为更广泛的犬主提供合适的犬只，还有助于在整个品种中维持更高的遗传多样性。
威胁：由于“拆分”策略旨在消除采用“趋同”或“分化”方法可能出现的问题，其风险远低于那些选项。通过允许和鼓励既繁殖原始品种用途犬，又培育新的友善伴侣类型，不存在失去传统工作功能或没有该品种城市兼容版本的危险。
6.4. 策略“合并”：创建新的城市犬种
这一途径可以被视为将先前一些策略的最佳可能结果融合起来，同时也允许繁育者和犬主选择更传统的解决方案。“趋同”策略的主要问题在于，繁育者和品种爱好者可能都不愿意将他们喜爱的品种改变为统一的“友善”城市犬。在“趋异”策略的情况下，由于其成功依赖于犬主在未必适合城市伴侣犬地位的犬种中做出有根据的选择，他们很有可能做出次优决策。“融合”策略强调需要新创建的品种，因为选择的主要目标将是满足一种随和、友善的城市伴侣犬的需求，这种犬很可能有几种不同的体型和外观（例如，被毛类型、耳朵位置等）。通过创建新品种，传统品种的爱好者不会觉得他们喜爱的犬的特征受到威胁，而那些只需要一个理想的城市伴侣犬、无需过多准备和训练的潜在犬主，也能找到一个直接的解决方案。
我们必须承认，在我们概述的四种策略中，“融合”可能不仅是最有前景的，也是最具挑战性的。那些在20世纪上半叶通过计划杂交和仔细的后续选择而创建的犬种，主要是根据特定的功能需求作为工作犬来培育的（例如，德国牧羊犬、莫斯科护卫犬、捷克斯洛伐克狼犬）。目前，一些受欢迎的工作犬品种之间有时会进行杂交，目的是通过更高的遗传多样性来维持强健的健康（例如，服务犬和导盲犬项目中的拉布拉多犬和金毛寻回犬：[227]）。据我们所知，迄今为止还没有任何犬种是以成为理想伴侣犬为目标而进行选择的。近年来呈指数级增长的“设计师杂交犬”库存的流行，更多显示的是其创造者作为营销者满足买家对不寻常外观偏好的聪明才智，而非对积极行为特征的任何规划[70]。因此，那些着手基于友善特征建立新品种的繁育者需要：(1) 找到最有前景的“起始”品种（类似于“趋同”策略）；然后(2) 创建品种杂交，并通过定向选择向最佳平衡的特征推进；最后(3) 稳定犬的特征。只有在新“品种”能够以合理稳定的行为和外观可复制之后，才能开始在潜在犬主中推广。
优势：采用这一策略，犬主可以获得功能性繁育的犬，这些犬将可靠地表现出理想的城市伴侣犬角色的行为特征。这些新类型将为传统上偏好的伴侣犬品种提供有利的替代选择，其优势在于拥有多个祖先品种，这可能带来更高的遗传多样性[228]。目前现有犬种数量众多，可以为初始选择提供广泛且足够多变的原材料，特别是如果有分子遗传学结果辅助识别最受追求（和最不受欢迎）的特征。关于人类导向社会行为的遗传力和遗传变异性，已经有令人鼓舞的结果（例如，[229,230]）。“融合”策略的另一个优势是，虽然它可能为市场增加新品种，但会使现有犬种保持其当前的行为多样性（类似于“趋异”策略）；因此，它不会引起品种爱好者的强烈抵制，因为他们不希望看到其品种的原始特征被改变。
劣势：这一策略可能最难在更大范围内建立。新“友善”品种的繁育者可能面临犬只注册的问题，因为主要的犬业俱乐部不接受此类尝试。虽然“趋同”和“趋异”策略不涉及杂交，因此产生的犬可以容易地注册为传统纯种犬，但“融合”策略的成功将在很大程度上取决于繁育者和犬主的热情、营销和教育。设计师杂交犬在犬种爱好者和繁育组织中并未得到一致欢迎（例如，[164]），这对专门培育的新“友善”品种也是一个警告信号。“融合”策略的另一个劣势是其缓慢的节奏。如果繁育者旨在创建“稳定”的新伴侣犬品种，可能需要几代人的精心选择（平均每代3年）才能使行为特征稳定下来。作为比较，捷克斯洛伐克狼犬在最初及随后与狼杂交后，花了不少于25年才稳定下来[231]。即使有最佳的开始和广泛的教育/营销项目，我们可以预期“融合”策略在很长一段时间内仍将是面向最忠实爱好者的一个小众“行家”选择。这将使当前选择不太适合的犬作为城市伴侣犬的问题基本保持原状。
机遇：通过提供以最适宜的友善伴侣犬特征为特点的全新犬种，“融合”策略相较其他三种策略更容易获得热情的支持者群体。新培育的犬种不会给人留下这样的印象，即针对城市伴侣犬的选育会危及传统犬种的原始品质。将教育犬类爱好者了解城市伴侣犬理想特征以及宣传负责任养犬的机会与新选育犬种的推广相结合，可以获得新的动力。当友善伴侣犬行为的好处与一种新型犬相关联时，比主张历史悠久的犬种应经历根本性的行为改变，更容易说服犬类爱好者。可以在新的友善犬类型的育种计划中使用混血犬，因为通过行为测试和遗传评估，那些具有某些理想品质的犬也可以成为有价值的育种来源。
威胁：该策略即使成功，也只有在很长时间之后才能小规模地获得新培育的伴侣犬。对备受追捧的理想友善犬的需求可能为无良繁殖者创造机会，提供来源可疑的低质量、繁殖不良的犬只，类似于当今流行犬种繁殖中的倾向[128]。当关于新友善伴侣犬种的积极经验和成功宣传导致购买价格高昂时，这一现象也会加剧。我们也不应忘记许多犬类爱好者的保守态度，使他们不愿改变自己的犬种偏好。以下事实充分证明了这一点：尽管多项实证结果表明极端短头犬种存在复杂的健康和福利受损状况（如[232]），但其忠实的饲养者群体仍拒绝停止选择这些犬种[214]。出于不同的原因，从收容所领养问题犬（主要是混血犬）的支持者可能也会抵制选择新的伴侣犬种。
表1显示了每种策略的SWOT评估的主要细节。
Table 1. The summary of the SWOT analysis in the case of each strategy proposed in this article.
| Converge | Diverge | Split | Merge | |
|---|---|---|---|---|
| Strength | As a result, widely available amicable urban dogs from many breeds. | The widest involvement of breeds and existing breed-related knowledge. | Flexible application to most dog breeds. Already existing within-breed lines. | Offering new amicable urban dogs and leaving the original behavioural traits intact in the ancestral breeds. |
| Weakness | Selection for ‘amicability’ would not be feasible in many breeds. Parallel selection for several preferred traits would be difficult. | Extensive need for educating the prospective owners. Often strong within-breed behavioural variability. | In the case of already low genetic variability, the dual selection purpose might not be feasible. | Slow progress is due to the difficulty of creating new ‘breeds’ with a targeted behavioural complex. Strong resistance from traditional breeders’ organizations. Possibly small-scale output. |
| Opportunity | ‘Last chance’ to save some rare or less popular breeds. Developing such behaviour tests that can also be used for assessing shelter dogs. | Widespread availability of reliable working dogs. With the right education, a safe choice for companion dog owners. | The highest acceptance among the various enthusiast groups. Good chance for retaining maximum genetic variability within the whole population. | The de novo created urban dog breeds would be suitable for effective education about the advantages of amicable dogs. Less friction with traditional breed enthusiasts. Easier involvement of mixed breeds for the breeding program. |
| Threat | Random loss of such, often-preferred behavioural phenotypes that characterized the original working dogs. Unpredictable emergence of unwanted new behavioural traits. | Disappearance of obsolete function breeds. Strong chance of loss of genetic diversity. Purpose-bred amicable companion dogs will be missing. | Not applicable, as the ‘Split’ strategy is intended to attenuate the threats that may affect the ‘Converge’ and ‘Diverge’ strategies. | High demand for the available new amicable dogs, and hence opportunistic suboptimal breeders’ practices. Ethical issues with potentially compromising the adoption of shelter dogs. |
| Main steps | Identify the most suitable breeds. Repeated rounds of directional selection and behavioural assessment. | Identify optimal breed functions. Repeated rounds of stabilizing selection and behavioural assessment. | Assessment of within-breed behavioural variability. Depending on the prevalence of ‘amicable’ and ‘working dog’ traits, applying disruptive selection to achieve both goals. | Behavioural and, ideally, genetic assessment of initial breeding animals to identify the most suitable dogs. Repeated rounds of directional selection and behavioural assessment. |
7. 讨论
伴侣犬在世界范围内享有前所未有的、且仍在稳步增长的人气[53]。除了人与狗互动不可否认的互利益处[233]之外，动物和人类两方面所产生问题的严重程度也令人担忧[181,234]。在许多潜在影响因素中，城市养狗比例不断增长起着关键作用。许多狗似乎不适合城市居住条件，也难以适应主人忙碌的生活方式，这伴随着长时间的隔离（例如[64]）。与此同时，随着养狗成为社会常态甚至近乎时尚（例如[235]），狗主人对狗的需求和自身责任的平均知识水平仍然不足[236]。在这篇观点文章中，我们提出一个发人深省的解决方案：我们能否通过重点培育以下类型的狗来改善上述突出问题：（1）更符合“友善城市伴侣”标准的狗（遵循“趋同”、“分离”或“合并”策略）；或（2）根据其原始品种功能获得更可靠、更可预测的行为特征的狗（“趋异”和“分离”策略）。
这些策略旨在主要通过利用人工选择的优势来改变狗的行为特征。显然，各种训练方法可以为影响大多数狗的行为提供更快、更易获得的工具（例如[237]）。我们并不声称通过选择最符合主人期望和条件的狗或狗品种，就能完全替代随后对狗进行训练和共同活动的需要。然而，有经验证据表明，不同狗品种对人类教导努力的反应强度（例如[43]）和效果（例如[42,194]）不同。被选择与人类执行合作任务的狗品种更热切、更好地从人类指导者那里学习[42]，它们更有效地跟随人类交流信号[79]；但它们也更容易出现分离相关问题[132]。此外，不仅是功能选择，与狗共同祖先的遗传距离也会影响狗对人类存在和交流的反应性。根据一项使用C-BARQ工具的问卷调查，原始（古代）狗品种被发现对主人的关注度较低，依恋较弱[238]；而基于对人类示范者的社会学习表现，现代猎犬（如寻回犬、指示犬、獚犬）和实用犬种（德国牧羊犬、霍瓦瓦特犬）最为专注且最依赖人类行为[25,239]。因此，我们可以得出结论，目前各种狗品种集群以及混种狗的巨大变异性，使得对于那些希望让这些狗适应“理想城市伴侣”要求的人来说，很难勾勒出一个普遍可靠且易于使用的训练方案。因此，我们认为，通过有目的的行为选择来创造更合适的狗（品种）的倡议，将是一个有用的起点，可以在城市中与我们的伴侣犬减少问题、更愉快地共处。
显然，如果没有狗繁殖者、最好还有国际繁殖组织的广泛协调努力，我们提出的任何策略都无法奏效。此类倡议已被建议用于补充高质量检测犬[91]；改善犬类健康、福祉和福利[240]，以及解决纯种犬遗传健康问题[241]。然而，关于为满足公众需求而改善狗行为的有组织努力，我们只知道少数相对小规模的干预措施。其中之一记录在荷兰，那里罗威纳犬需要培育出低水平的攻击性和恐惧性，以避免立法禁令。据报道，当只有亲代狗成功通过官方行为测试才能登记罗威纳犬窝时，最终荷兰罗威纳犬开始在攻击性和恐惧性指数上表现出改善[200]。针对更广泛的行为特征进行育种，例如“友善伴侣犬”的要求，不仅需要明确、基于共识的目标行为和合适的评估测试方法，还需要许多敬业的繁殖者与一个协调委员会保持一致，该委员会将提供必要指导并促进繁殖者之间的合作。对于“趋同”和“合并”策略尤其如此。后者旨在开发新的友善城市狗品种，而前者鼓励参与的繁殖者在现有狗品种内进行定向选择以趋向友善特征。在这些情况下，成功将取决于对所创造的新一代狗进行适当和标准化的行为评估。“趋异”和“分离”策略会更容易，因为每个涉及的品种（或在“分离”情况下品种的亚群）都需要围绕其标准中描述的、并由实践证据支持的典型功能行为特征进行稳定。
新西伯利亚农场狐狸实验表明，通过对毛皮狐狸（即“野生”动物）种群施加严格的方向性选择，始终使用每一代中最温顺的10%进行后续繁殖，可以相对快速地朝着理想的行为表型改变[242]。在我们提出的策略中，如果育种者仅选择表型最佳的个体进行后续繁殖，也可以以合理的速度培育出友善的伴侣犬。如果所需性状的遗传背景已知，该过程还可以通过基因检测加以辅助。然而，将较大比例的后代排除在繁殖计划之外需要行政措施，即那些行为性状不合适的犬随后不能用于繁殖。犬的繁殖在很大程度上是一种休闲爱好，其组织化、规范化和商业化程度远低于通常的牲畜繁殖（例如[243]）。因此，任何提出的犬行为改良策略都需要大量的公众支持，以及来自科学界和兽医的专业指导；最重要的是，犬繁殖组织和育种者本身的兴趣和热情是必不可少的。
每种策略都将利用混血犬并对其产生影响，混血犬不仅数量众多，而且是全球伴侣犬群体中广受欢迎的成员。获取混血犬的动机可能包括：(1) 经济原因；(2) 希望从收容所领养犬只[244]；(3) 认为混血犬比纯种犬更健康[70]。有实证证据表明，混血犬至少与纯种犬一样可训练且专注[121,245]；然而，也有研究表明混血犬表现出更多问题行为[121]，且阅读人类视觉信号的能力不如纯种犬[79]。由于对犬进行理想“友善”性状的行为评估将是“聚合”、“分离”和“合并”策略中不可或缺的工具，混血犬也可以接受这一程序。表现良好的个体可以被标记为理想城市伴侣角色的高度合适候选者，或在“合并”策略原则下用于开发新的犬品种。同时，由于所提出的策略主要旨在选育改良的纯种犬（尤其是在“聚合”、“分化”和“分离”策略的情况下），推广工作也将主要针对这些而非混血犬。希望对友善伴侣犬的显著需求也能间接强调混血犬领养前行为测试的重要性[246]。
最后，我们强调犬主自身在每项提出的伴侣犬改良策略中的关键重要性。潜在犬主代表了犬只生产和供应至“市场”的最终驱动力，其中育种者和收容所作为供应方。如果新犬寻求者对可靠表现出友善城市伴侣性状的犬只施加其有见识的需求，信息将清晰地传达给供应方。当前数据显示，犬主获取犬只的原因呈现混合图景：虽然犬的外观和行为对他们很重要（例如[80]），但同样经常地，他们的选择基于情感原因或一时冲动（例如[82,133]）。在我们提出的策略中，犬主需要理解犬的“友善性”是在城市环境中更愉快共存的关键组成部分。通过“聚合”和“分离”策略，我们可以提供广泛的合适犬品种选择来满足这一需求。在“分化”和“分离”框架下，犬品种将更可靠地展现其原始行为性状，因此受过良好教育的犬主可以安全地选择那些最能展现理想城市伴侣特征的品种。如果采用“合并”策略，全新的一类专门培育的伴侣品种可以服务于有知识的犬主的需求。
犬主教育项目广为人知，并因各种原因日益被采用。已有倡议教育犬主如何更好地确保犬的一般健康和福利[212,247]、对抗犬肥胖[248]或管理攻击行为[249]。最近，已投入大量努力向极端短头犬爱好者解释这些品种的不良福利前景[250]。由于对易于共处的城市伴侣犬的需求基于对犬主的实际调查[64,65,185]，我们可以对传播四种策略中任何一种积极特征的宣传活动潜在成功持乐观态度，这些策略均针对培育合适的友善伴侣犬。与极端短头犬的情况不同——在这些情况下，犬主对其喜爱品种的偏好迄今已被证明对动物福利不佳的严峻现实具有抵抗力（例如[214,250]）——城市中对友善犬的需求反映了公众需求，这可以为此类倡议取得更大成功带来希望。
8. 局限性与潜在困难
在本文中，我们提出并评估了一些观点，我们认为这些观点可以为有组织的育种工作提供连贯的指导方针，从而培育出更适合城市伴侣用途的犬只。在某些情况下（“分化”和“拆分”策略），这些策略甚至可能提供能够可靠展现其原始功能适应性行为特征的工作犬。然而，这些策略的广泛实践成功在很大程度上取决于许多因素，而且所提出的育种策略也可能引发伦理担忧。我们在此简要提及其中最相关的问题。
8.1. 犬行为的遗传背景
尽管有一些令人鼓舞的例子表明，针对所需行为特征的方向性选择取得了显著成功（例如，荷兰罗威纳犬攻击性的降低：[200]），但据我们目前所知，大多数分子遗传学研究已检测到犬行为表型背后的多基因背景[251,252]和广泛的基因多效性[95,151,208]。当育种者希望选择更复杂的行为特征——这些特征在“理想的友善城市犬”中更为可取——时，这构成了严重的困难。
8.2. 大规模犬类育种计划的困难
由多基因行为特征、低遗传力和环境影响所产生的问题，或许可以通过使用前沿分子遗传学、大量繁殖动物、集中化育种计划、严格的选择标准和标准化环境来克服——正如我们在畜牧科学中所见（例如，[253]）。除少数例外，犬类育种是一项休闲活动，前述大多数条件均不具备。有希望的是，犬类分子遗传学研究的进展将为更有针对性的选择提供指导。然而，要成功实施大规模育种计划，需要众多热心育种者的自愿联合，并得到那些接受所得犬只登记的进步育种组织（犬业俱乐部）的协助。此类努力的资金背景也是一个关键因素：许多非商业性犬类育种者需要为共同育种项目的初始阶段提供资金；同时，所得犬只的预期有益特征应在潜在客户中加以推广。
8.3. 遗传变异性的潜在丧失
由于大多数犬种的血统登记簿是封闭的，与品种典型的原生地方犬种或其他犬种进行异系交配实际上是不可能的。这一点，加上不适应的育种实践（例如，使用“流行种公犬”：[225]），已导致许多犬种遗传多样性的显著丧失。由于我们提出的一些育种策略必然包括针对所需行为表型的方向性选择，这可能进一步导致所得犬只遗传变异性的丧失。这需要多个育种者的协调努力，他们可以维持具有所需重点特征的遗传上独立的育种品系。
8.4. 关于收容犬群体的伦理考量
本文提出的育种策略侧重于改良犬种，以提供行为可预测的、适合城市的伴侣犬。“合并”策略甚至勾勒了通过行为评估、定向杂交和后续选择来创造新的友善伴侣犬品种的构想。然而，我们不应忘记，全世界有数十万只犬正在收容所、动物庇护所和各种救援组织中等待领养（例如，[254,255]）。尽管相当数量的犬只因行为与主人期望不匹配而被弃养[40]，且收容所中犬只缺乏适当的行为评估可能带来进一步的问题[256]，但我们的道德义务是优先领养那些被安置在救援组织和收容所中的犬只，这些犬只能够安全可靠地重新融入伴侣犬群体。由于对偏好的“友善”特征进行行为测试是我们所提出育种策略的工具性要素，这些方法也可以应用于收容所中犬只的行为评估，从而提高它们被领养的机会。
9. 结论
我们与犬的关系是家养动物中最古老的关联，并且在过去的几千年中经历了持续的变化。犬的形态和功能随着与其共同生活的人类的需求、偏好和可能性而平行演变。然而，自19世纪有组织的犬类育种出现以来，人为制造的品种标准化需求似乎一直在（并且仍在试图）阻碍犬只根据社会和个体犬主的新要求和需求进行持续发展和转变（即“适应”）。在世界人口最稠密地区犬只饲养空前普及的背景下，我们现在的任务是为未来新一轮犬种的选育和培育提供专家指导。犬种的未来将取决于它们能否成功适应我们的城市生活方式以及对其原始功能的需求，而我们有责任在塑造它们的行为特征时兼顾这些特征对犬和人的益处。
为总结本文的主要信息和结论，我们想强调以下几点：
（1）城市养犬者更偏好“友善伴侣犬”，即无攻击性、对儿童可靠、能量水平低、可长时间独处的犬。
（2）大多数犬种最初是为执行工作任务而选育的，因此作为城市伴侣犬会表现出若干不理想行为；同时，混血犬的行为特征可预测性较低。然而，我们推测，通过对现有犬和未来繁育犬进行适当的行为测试，在大多数犬种中通过遗传选择获得理想的“友善犬”特征是可能的。
（3）我们提出四种策略，通过有目的的繁育创建新的犬种类型，目标是提供可靠、安全的伴侣犬。这些犬有望满足主人对友善伴侣动物的需求，而且由于其问题行为发生率较低，也更适合城市生活，从而提升其福利。
（4）策略“趋同”将鼓励犬繁育者在大多数现有犬种中对“友善”特征进行定向选择。因此，养犬者可以获得许多具有适合城市生活的理想行为特征的合适犬种。该策略尤其适合那些已表现出有前景的伴侣犬行为特征的犬种，以及那些未表现出或仅表现出少数严重不理想特征的犬种。
（5）策略“趋异”将强化现有犬种原有的行为特征；因此，主人可以更容易地识别出最适合城市生活的犬种，而这些犬将可靠地表现出其犬种所期望的行为特征。该策略尤其建议用于那些仍常被用于执行其原始任务或大量参与专业犬类运动的犬种。此外，那些主要表现出不适合作为友善城市伴侣犬的不理想行为特征的犬种，也将从“趋异”方法中获益更多。
（6）策略“拆分”在犬种内结合了“趋同”和“趋异”；因此，它有助于创建一个伴侣犬品系，同时保持一个可靠的工作品系。该策略对于既需要作为伴侣犬又需要作为专业工作犬的犬种非常有用；然而，这两种功能需要明显不同的行为特征。
（7）策略“合并”将启动创建表现出理想城市伴侣特征的新犬种。在选择时，可以利用现有犬种的杂交，以及具有偏好特征的混血犬。这可能是最有趣且最有前景的策略，因为它或许能够最有效地推广。然而，由于救助组织有大量可领养的犬只，创建新的伴侣犬种应在伦理上具有合理性。
（8）任何这些策略的成功，都需要许多热心的繁育者、繁育组织和专家协调委员会的支持。
（9）由于对随和、友善的城市伴侣犬的需求基于养犬者的实际需求，我们相信通过这些策略产生的新犬可能会广泛流行。然而，要实现这一点，无论繁育者选择哪种策略，都需要为潜在的犬类爱好者提供广泛且传播良好的教育项目，最好从早期年龄开始。
致谢
作者感谢Celeste R. Pongrácz对稿件的校对。
作者贡献
概念化：P.P.、P.D.；调查：P.P.、P.D.；方法学：P.P.；项目管理：P.P.；监督：P.P.；可视化：P.P.、P.D.；写作——初稿：P.P.、P.D.；写作——审阅与编辑：P.P.、P.D.。所有作者均已阅读并同意稿件的发表版本。
机构审查委员会声明
不适用。
知情同意声明
不适用。
数据可用性声明
本研究未创建或分析新数据。数据共享不适用于本文。
利益冲突
作者声明无利益冲突。
资助声明
Péter Pongrácz和Petra Dobos得到匈牙利国家研究、发展与创新办公室（NKFIH，资助号K143077）和Morris Animal Foundation（D25CA-039）的支持。Petra Dobos得到创新与技术部大学研究奖学金（EKÖP-25-3-I-ELTE-147；EKÖP-26-3-I-ELTE-2026012）的支持。
脚注
[内容请查看原文]
参考文献
[内容请查看原文]
关联数据
数据可用性声明
本研究未创建或分析新数据。数据共享不适用于本文。
查看英文原文▾
1. Introduction
Nobody Now Who Is Anybody Can Afford to Be Followed About by a Mongrel Dog
This famous line, quoted from the prolific 19th-century writer William Gordon Stables in 1877, and cited from Ritvo [1], provides the perfect prelude to the topic of this paper. The Victorian era saw an unprecedented popularity explosion of leisure dog keeping among noblemen and middle-class urban citizens in Great Britain [2], and the fancy was quickly copied and adopted all around the industrialized world [3,4]. In our opinion, finding a solution to many problems that today’s dog owners are facing regarding the optimal and ethical adjustment of dogs according to the expectations of the surrounding society would be impossible without understanding how and why the concept of modern dog breeds was born. Dog owners and breeders can only work with those dogs that are the products of the organic development of pedigree dog breeding over the previous 150 years, which shaped not only their look but also defined the behavioural characteristics of the dog breeds.
Importantly, the Victorian fashion of companion dog ownership focused on a basically new ‘invention’: the purebred dog [5]. The standardization of dog breeds, known mainly as ‘land races’, became a new source of national pride [6,7]. By the beginning of the 20th century, urban (companion) dog keeping was already a popular hobby [8]. From the rural working dog types (livestock guarding, herding and multi-purpose farm dogs), and hunting dogs of noblemen, which were mostly unknown to the public a century before [9], a new genre of utility/service dogs emerged from the cataclysm of the First World War [10]. These ‘modern working dogs’ quickly became enormously popular, among the public and various armed forces, opening new avenues towards a more versatile utilization of dogs as working and sports companions for humans (e.g., [11,12]).
Studying dogs’ behaviour has become increasingly popular since the 1990s [13]. The main breakthrough for ethologists came from the notion that the natural environment for dogs is the anthropogenic niche [14,15], because this concept directed the researchers’ focus on those socio-cognitive and behavioural capacities of dogs that enabled them to adapt to the human social sphere as successful commensalists [16,17] and companions [18,19]. A systematic summary of the vastly complex field of scientific research related to dog–human interactions is beyond the scope of this paper. Among the main directions is research regarding the evolutionary origins of dogs based on molecular genetics [20,21], where comparative behavioural research, including the closest wild relative, the grey wolf (Canis lupus), may shed light on the possible origins of characteristic human-related socio-cognitive features of the dog [22,23,24]. A considerable segment of empirical research focuses on the role of artificial selection (ancestry: [25]; functional selection: [26]; working lines: [27]) and environmental effects (upbringing: [28]; training: [29]) on the behavioural variance among various cohorts of dogs. A quickly growing field of research uses dogs as a promising, biologically relevant model species for numerous human disorders at the level of genetics (e.g., [30]), physiology (e.g., [31]), and behaviour (e.g., [32]). Finally, the characteristics and newly emerging problems of dog–human relationships provide ample opportunity for applied ethological research as well (e.g., [33,34]). Problematic behaviours in dogs have a very complex and multi-faceted background, where the environment (e.g., socialisation: [35]; training: [36]; trauma: [37]) and the genetic background of the dog (e.g., [38]) can both influence the behavioural phenotype. In this paper, we are going to focus on those potential risk factors that stem from the mismatch between the originally sound breed-typical behaviour/temperament, the owners’ expectations, and environmental conditions (e.g., [39,40]).
Breed standards and breeding organizations still proudly refer to the original functions of the dog breeds (e.g., [41]), but new-age urban dog owners do not usually rely on those aptitudes and function-specific behavioural traits that the various working dog types have been selected for through countless generations. Although the strict isolationist principles of pedigree dog breeding effectively retained many of the main behavioural characteristics of functional working breed selection (e.g., the differences between independent and cooperative working dog breeds: [42,43]), the detectable division of many dog breeds into ‘show’ and ‘working’ lines (e.g., [44,45]), signifies the redirected aspects of directional selection compared to the traditional utilitarian methods of dog breeding.
It is safe to say that towards the 21st century, companion dog ownership became a lifestyle [46]. According to estimations, more than two hundred million dogs are being kept as pets around the world [47]. Furthermore, based on several questionnaire studies, modern dog owners increasingly regard their canine companions as family members, or even as ‘children’, instead of referring to them as simply domestic animals (e.g., [48,49]). The popularity of dog keeping has reached such levels in industrialized countries that some scientists have even started hypothesizing that companion dogs may act as surrogates for their owners, instead of them having their own children [50]. Recent research provided more elaborate detail on this matter, showing that companion dogs offered women more social flexibility and connectedness to the social environment than children; however, mothers reported higher levels of responsibility towards their children than their dogs [51]. Thus, although the human–dog emotional bond can be comparably complex and as deep as a human–human emotional relationship, the social status of companion dogs is also influenced by the family structure of their owners.
Urbanisation, a worldwide process where the density of human inhabitants shows an ever-accelerating ratio, brings new challenges for dog–human coexistence (e.g., [46]). As in ‘Western civilization’, most owned dogs are kept as companions (e.g., [52,53]), urban inhabitants demand the right to keep dogs, even if the traditional working roles of these animals cannot be fulfilled in the metropolises. Recently, city planning and human environmental research have started to assess the conditions, possibilities, and limitations of urban dog keeping among the many issues to be considered (e.g., [54,55]). While the natural needs of dogs for activity and environmental enrichment are easier to fulfil if they are kept at least partly outdoors (e.g., [56]), urban dog keeping poses serious challenges for dog owners. It is not surprising that urban planners, legislators and animal welfare experts pay close attention to many aspects of the ‘best practices’ of urban dog ownership, from the minimum number of daily walks to the maximum duration of being left alone [57], how to handle dog-derived waste [58], and the issue of nuisance barking [59].
As the behaviour of dogs represents perhaps the main synchronizing factor towards a mutually enjoyable coexistence between humans and dogs [19], many aspects of dog behaviour receive significant attention regarding their suitability for urban dog keeping. It is important to see that such characteristics that were typical and even preferred in various working dog breeds may be regarded today as unwanted for a city companion. Although nuanced details of what is an ‘ideal dog’ may vary among countries with different cultural, economic, and historical heritage [60], high activity levels, protection and guarding, being an alert and loud watchdog, and being inclined to chase and hunt are just a few examples of traits that are rarely welcome in an urban companion dog.
The recent surge of (urban) dog keeping has noticeably resulted in a novel set of demands from the ‘end users’ (dog owners), as well as from society regarding the behavioural characteristics of dogs [61]. While throughout the 20th century, advice regarding the suitability of one or the other dog breed for a particular owner remained within the scope of matching the size, temperament, and perhaps the utility of the breeds to the needs and lifestyle of the owner (e.g., [62,63]), recently, more specific expectations have been outlined that imply the need for a brand-new type of dog. The repeatedly surveyed opinion of Australian dog owners about the traits they prefer in ‘the ideal companion dog’ [64,65], crystallized the image of the ‘amicable’ dog type. This can be characterized as a low-drive, healthy, child-safe, non-barky, and non-aggressive animal that, regardless of the breed, should be the perfect companion for urban-style dog keeping. Although some subcultures still prefer characteristically non-amicable dogs, and use aggressive, menacing-looking dogs as status symbols [66], the worldwide dominant trend is leaning towards friendly and safe urban companions. Based on this, one can assume that the scene is finally set for a new era of directional selection for dog breeds that meet the demands of the largest population of prospective dog owners. Our goal with this paper is to critically assess the potential solutions to this paradoxical situation, where the traditional conservationist world of purebred dogs faces probably the biggest challenge since its origins one and a half centuries ago.
2. Methods: Literature Survey
We did not aim to write a systematic review; however, to better substantiate the background of our opinion piece, we ran a literature search between 20 and 23 May 2025. We explored Google Scholar and ScienceDirect using four sets of keywords: “dog breeds AND future”; “expectations AND dog breeds”; “change AND dog breeds”; and “improving dog breeds”. An additional literature search was run between 14 and 16 August 2026, on Google Scholar and ScienceDirect, where we used two new sets of keywords: “urbanisation AND dog keeping AND demands”; and “urban dog keeping AND problematic behavio(u)rs”. For each search, we examined the titles and abstracts of the first 100 hits. Based on our experience with previous literature surveys, including more hits per search resulted primarily in duplicates and thematically unsuitable items. We excluded non-peer-reviewed items, dissertations, and conference abstracts, and we kept monographs, book chapters, and peer-reviewed journal papers for further assessment. Our primary interest was in dog behaviour, breed history and dog–human interactions, with an additional focus on the specific issues relating to urban dog keeping. The final set of retained papers comprised N = 60 publications after excluding all duplicates and out-of-scope items. We further enriched the literature collection by examining the cited references in some of the best-fitting articles from our primary search.
3. Dog Types, Dog Breeds and Associated Behavioural Traits: A Brief Introduction to Ancestry-Based and Functional Selection
According to our current understanding, dog breeds are defined by and distinguished from each other based on three separate dimensions: (i) their ancestry (i.e., purebred dogs must be the descendants of a known and isolated group of founders); (ii) their function (i.e., original purpose-driven behaviour); and (iii) their typology/look [41,67]. It is important to understand that these three components became an indispensable part of purebred dog breeding only during the Victorian era. Closed pedigrees and the strict parameters of a dog’s appearance, as defined by breed standards, did not exist before the 19th century, and the formation of the various working dog types (landraces) was mainly driven by behaviour-based functional selection (e.g., [3]). Since then, the picture has completely changed. Although the alleged original functions of dog breeds are still highlighted in the nomenclatures of leading kennel club organizations [41], modern dog breeders put less emphasis on function-specific behaviours, which leads to increasingly uniform, uncharacteristic behavioural traits in various breeds bred mainly for conformation shows [68] and canine sports. In return, instead of the original strong directional selection towards desirable function-specific behaviours, an even more pertinent selection for the (often extreme) visual appearance of dog breeds started to characterize the last century of the dog fanciers’ world [69,70].
Although the traditional functional selection pressure, which resulted in the multitude of working-task-specific dog types (and eventually breeds), has been mostly relaxed in present times, it is easy to see that the main behavioural/temperamental traits of dog breeds can still show striking differences. Recently, a large-scale molecular genetics study could not detect breed-specific behavioural phenotypes [71], but the same paper highlighted several important genetically determined features (e.g., ‘biddability’) that overarch several dog breeds, placing them along a scale of intensity of this behavioural trait. These findings warrant the need to shift the focus from making breed-level behavioural comparisons (e.g., [72,73,74]), towards finding and investigating broader explanatory factors that can influence the socio-cognitive and behavioural phenotypes of larger clusters of breeds [41,75].
Ecological validity (i.e., why particular phenotypes would be advantageous to the investigated subjects in a given context) should always be kept in focus when we want to understand why specific behavioural phenotypes emerged in some groups of dogs, therefore functional breed selection (e.g., [43,76]) offers an especially suitable approach for testing biologically relevant differences among clusters of dog breeds. We should note that dogs’ behaviour, and particularly their social interactions with humans and other living beings, is of decisive importance in the assessment of existing dogs [40,77] and ‘engineering’ dogs for the future (e.g., [64,78]). Artificial selection efforts that shaped the various dog breed types could potentially influence fundamental socio-cognitive capacities among working dogs, depending on their intended usage. A comparison between ‘independent’ and ‘cooperative’ working dogs [79] is based on the distinct interaction styles of these dogs with their owners. Cooperative breeds originally performed their job under close human guidance with regular instructions (e.g., gundogs, herding dogs). On the other hand, hounds, livestock guardian dogs (LGDs), and terriers have been selected for tasks in which they must operate independently, either out of sight of their handlers or without being regularly instructed. From the aspect of looking for suitable dogs for modern ‘end users’ (i.e., urban companion dog owners), the functional clustering of dog breeds has a specific advantage. Cooperative and independent dog types include a wide variety of breeds, with highly variable morphological characteristics. As physical appearance is among the most important factors of preference when choosing a companion dog (e.g., [80]), the variable breed assortments with predictable behavioural characteristics within the cooperative and independent dog breed types have considerable value for prospective dog adopters. Trainability [81,82] and interactivity of dogs [83,84] are important assets for companion dog owners. In connection with these characteristics, it was found that functional breed selection affected those socio-cognitive capacities of dogs that could also play a role in everyday dog–human interactions. Cooperative dog breeds established eye-contact more easily with humans [85], they followed human-provided visual pointing more successfully [79] and they also learned better via the observation of a human demonstrator [42]. Meanwhile, it was found that independent breeds performed fewer gaze alternations when facing a difficult task [86], were more likely to steal forbidden treats when owners could not see them [87], and in general, they paid less attention to human verbal communication than cooperative breeds did [43].
Artificial selection for various working tasks in dogs has not ceased; it may have just taken different directions recently, for example: selecting dogs for novel functions such as guiding people with visual impairments [88]; serving as assistance dogs [89]; or detecting the smell of explosives [90,91]. However, the divergence between the various dog breed types has not been solely driven by the breeding efforts of humans. Based on molecular genetics, today’s dog breeds (as well as the free-ranging and feral dogs) [92,93] can be effectively clustered on the basis of their genetic distance from the common ancestor species [94,95]. Besides the functional selection-based clustering of dog breeds, ancestry-related associations between breed clades and specific behavioural traits represent another main, biologically relevant approach for behavioural investigations [41]. Regarding the main topic of this paper, we should see that the genetic distance from the hypothetical ancestor of all dogs aligns well with some of the most important socio-cognitive and behavioural characteristics of dog breeds from the aspect of their suitability to the role of a companion animal. For example, ancient breeds (such as the Chow Chow, Akita, various sled dog breeds, etc.) showed lower trainability [96] and higher stranger-directed fear [97] than the more derived (‘modern’) breed clades. Ancient breeds also showed a more consistent relationship between sociability and reactivity [23]. According to the authors, this result could be explained by the domestication syndrome (e.g., [98]), because these breeds remained relatively unaltered for a long time, they show a more uniform effect of domestication. In the more derived breeds, the two traits are more separated, which highlights that in the modern breeds [94] sociability and reactivity had different importance during selection for special working needs. Besides individual behavioural traits, the more complex problem-solving ability of breed groups can be compared based on their ancestry. By observing a human demonstrator, the modern breeds, especially utility dogs, showed higher performance in social learning than the more ancient breeds that probably were not subjected to strong directional selection for showing high interactivity with humans [25]. Obviously, when one intends to draw valid and highly predictive conclusions regarding ancestry-based behavioural characteristics of dog breeds, the broad representation of breeds and an effective comparison among more distantly related breed clades remains a vital necessity. For example, when in a survey scientists studied the trainability of only ‘ancient’ dog breeds, their alleged low performance is nearly impossible to interpret without the possibility of comparison with more derived dog breeds [99].
4. Behaviours That Fit, and Those That Don’t: Preferred and Problematic Behaviours and Their Association with Breeds/Types
If we consider that there are more than 400 different dog breeds recognized by the main cynological organizations, it is not an easy task to decide which breed to pick for one’s next canine companion. According to surveys, the predictable behaviour of each breed or breed type is among the most important features to consider when prospective owners choose a dog (e.g., [80]). Although dog breed descriptions (‘standards’) primarily emphasize the characteristics of the breed’s physical appearance, they usually also list the breed’s original (work-related) function and typical behaviour. However, companion dog owners, veterinarians, and kennel clubs might have very different priorities regarding favourable and suboptimal traits and aspects of breed behaviour [11,44].
For instance, in Australian society, the ideal dog could be described as safe with children, easily house-trained, healthy, obedient, and of medium size [64]. That study found differences between the preferences of owners and non-owners, as well as between genders: women preferred a calmer, non-aggressive dog, while men opted for a faithful and energetic companion. These preferences were found to be relatively stable over time (see the repeated study from Australia [65]). On the other hand, the strong fluctuation of popularity among the various breeds did not show a connection with either the consumers’ canine behavioural preferences [100] or breed characteristics such as health, longevity, aggressiveness, trainability, and fearfulness [101]. Trends of breed popularity might be affected by different types of information during different historical periods (e.g., dogs as movie stars [8], appearing in TV commercials [102], or being associated with celebrities [103], etc.).
Dog breeds were mainly developed to perform some type of function [104]. While this seems obvious regarding working dogs (e.g., herding dogs) and modern utility breeds (such as guide, police or scent detection dogs, etc.), even toy breeds were selected based on functional aspects: to provide small and easily portable companion breeds for everyday life [105]. From this perspective, it can be assumed that behaviours that characterize a breed or group of similar breeds, might have been advantageous for their original function. Therefore, in the past, humans most probably put strong directional selection pressure on these once appropriate traits [5]. The vigilant, independent, and aggressive nature of a livestock guardian dog was needed to protect the herd from thieves and predators [106]; the high activity level and boldness of terriers were preferred for dispatching pests that foraged around the hen house [107]; and sled dogs could be independent and tireless as long as they performed well in front of a sled [108].
However, these specific features that were preferred for different types of work do not necessarily correspond to the lifestyle requirements of an urban companion dog (e.g., hunting behaviour that poses a threat to other pets, wildlife, and even small children: [109]). As soon as dog ownership became a leisure activity without the need for work-related functional use, the frequent confrontation with problematic behaviours started, mainly because the behaviour and temperament traits associated with the original working function did not always suit a companion. This process is well documented through the centuries-long records from London (UK), showing how the once useful watchdogs became today’s nuisance barkers [54,110]. Of course, there might be such breed types that make it easier to become a companion dog with predominantly amicable behaviour (e.g., based on the assessment of the complex preferences of dog owners from Australia: [65]). However, there might also be such breed types that are much less suitable for becoming companions in a family environment. Problematic behaviours usually have a complex causative background (e.g., separation-related problems, [111]). Thus, the endeavour to disentangle these from each other usually requires a large sample of dogs. For instance, it was found that one main grouping factor behind problematic behaviour was the size of the dog [112]. The authors discovered that each problematic behaviour, except low trainability, occurred more frequently the smaller the dog breed was. However, dog size can be confounded with other factors such as function, keeping conditions, and longevity [113]. Hence, many of these findings would still require further investigation.
Among problematic behaviours, (the lack of) aggression towards humans is one of the most important characteristics considered by owners when they choose a companion dog [82]. For example, terriers, especially the Staffordshire Bull Terrier, are viewed as even more dangerous in the opinion of the public than they are rated by legislators [114]. These often anecdotal perceptions of some dog breeds by society would need to be further confirmed by experimental evidence. In a questionnaire study, Asp and colleagues found that dog breeds from the AKC’s working dog group showed more aggressive behaviour than the non-working breeds [75]. On the other hand, based on empirical studies, there can be a large individual variation in aggressiveness and reactivity within breeds, regardless of whether the breed is subject to legislation [115], which is a common reason why people might oppose the ban of certain dog breeds [116].
Behind both the preferred and undesirable traits, there can be additional confounding factors. More fearful dogs were found to also be more aggressive [75], which could explain some of the individual and breed-related variance of this trait. In another questionnaire study, 17 breeds were found to have significantly different noise sensitivity levels [117], however, confounders such as the dogs’ age, sex, and sensitivity to separation-related problems were also found to have an effect. The heritability of these traits should be carefully studied to determine which are the least aggressive or fearful individuals for further breeding. However, due to the complex genetic background of most behavioural traits, it might be a difficult and lengthy mission. Even if researchers find genetic associations in one breed (e.g., fearful behaviour in the Bichon Havanese, [118]), the generalization of these findings might fail when implemented in different dog breeds [118].
Proneness to aggression is a crucially important trait when the adoptability of a dog is being considered. A brief encounter with a misbehaving dog before making the decision to adopt can unjustly cause a disadvantage for dogs of the same breed [119]. Unsatisfied owners of adopted dogs might return the dog to the shelter (even without concrete evidence of problematic behaviours [120]). In turn, purebred dogs are more expensive, and owners would rather live with the inconvenience [121,122], with or without attempts to improve the quality of coexistence through training. For mixed breeds, preadoption counselling with dog experts might be a good solution to increase the success of adoption [123], as it was found that dog-relinquishing owners had higher expectations regarding desirable behaviour and their dog–human bond [40]. Moreover, behavioural assessment of dogs in the shelter can greatly improve the adoptability of dogs [124] and potentially lessen safety hazards for humans.
Problematic behaviours may also occur due to the lack of strict behavioural selection during conformation show and pet dog breeding. While in utility dog breeding, high emphasis has been put on enhancing the efficacy of selection [27], the same cannot be stated in the case of dog breeds that lack a working function [125]. Despite the increasing interest and shift in breeders’ perception of breeds [126], it has become increasingly unpredictable what kind of behaviour and temperament can be expected from a particular breed. This tendency is supported by the attitudes of some breeders who, especially in the case of highly fashionable dogs, maintain and breed further animals with well-documented, chronic health and behavioural problems (e.g., ‘trancing’ behaviour in Bull Terriers: [127]; respiratory problems in brachycephalic dogs: [128]). The high inbreeding level of some dog breeds, or breeds with low genetic variability, is not necessarily beneficial to behavioural stability either [129]. However, high levels of inbreeding can be found in breeds considered ‘healthy’ and ‘unhealthy’ as well [130], therefore the exact behavioural- and health-related consequences of inbreeding should be disentangled with targeted investigations. For instance, it was found that inbreeding can affect dog–human interaction-related socio-cognitive traits, such as aversion to humans in pointer breeds [131].
Finally, a considerable number of reportedly problematic dog behaviours can be the consequence of the owners’ expectations being unrealistic regarding preferred dog behaviours. Many people want a human-centric, strongly attached canine companion [64], however, when they leave the dog during the day, they also expect that this highly human-dependent dog should not show separation-related behaviours [132] and should not bark excessively when they return home. On the other hand, owners who prefer a healthy dog continue choosing breeds with well-known heritable diseases and anatomical problems [82,133]. How to correct and resolve these unrealistic expectations requires initiatives to help owners recognize the importance of proper breed choice (e.g., [134]).
5. Multi-Level Attempts to Influence/Improve Dog Behaviour and Owner-Dog Matching Success: Line Selection, Breed Crosses, Puppy Tests
Although the Victorian-era standardization of pedigree dog breeding shifted the focus towards the appearance of breeds [1], the behaviour of dogs has never become of secondary importance [75]. This is reflected not only in the continued utilization of working dogs in various traditional (e.g., livestock guarding [135] and herding [136]) and newly invented jobs (e.g., various assistance [89] and service dogs [137]), but more importantly, in the gradual and most widespread role of being a pleasant animal companion, which required the constant maintenance of behavioural soundness [78].
When we state that the natural habitat of dogs is the anthropogenic niche [18], this also means that dogs must adapt to the emerging challenges of the human environment [19]. Of course, especially in the case of purebred dogs, adaptation happens mainly via artificial selection (e.g., [138], but principally the result is the same compared to animal populations under natural selection: traits of morphology and behaviour of the breeds are being shaped according to a more optimal fit to the human-imposed requirements [41]. This remains true, no matter whether we consider the need for suitable working dogs for specific purposes [27] or for the role of being a compatible pet. Companion dog owners have explicit needs according to their lifestyle [139], fashion [102] and environmental constraints [80], but dogs’ behaviour will always remain perhaps the most important phenotype that ultimately decides whether the match between the dog and the human is going to result in success or failure (e.g., [40,123]). It is no surprise that there are plenty of methods that people have invented and continue to use with the goal of improving or securing desired behavioural traits in dogs that would make them fit for specific purposes. These procedures range from directional selection for the desired (or against unwanted) behavioural traits through planned hybridization of distinct dog breeds to interventions that target the individual dog’s behaviour via early evaluation and shorter/longer training regimes.
5.1. Should I Stay or Should I… Disappear—The Potential for Directional Selection of Favourable Dog Behaviours
In breed standards, dogs are typically characterized by those behavioural traits that are considered ‘type-appropriate’ and preferred according to their original purpose. Curiously enough, of all the dog behaviours that attracted scientific interest, the unwanted, or even ‘abnormal’ phenotypes were among the first that were considered to have an inherited background. Among these, we can mention the ‘unstable temperament’ of Pointers [140], or the unprovoked aggression (‘spaniel rage’) in red/golden English Cocker Spaniels (e.g., [141]). Since then, and with only a few exceptions [64,78], the context of inherited behaviours in companion dogs has still been dominated by research efforts on various undesired phenotypes [142], while applied ethologists are more likely to focus on the preferred behavioural traits of utility/working dogs [143].
In a study, based on a database of results from many thousands of dogs from the Swedish BPH test (Behaviour and Personality Description in Dogs), Tygesen et al. [144] found five behavioural factors with a weak-to-moderate genetic background: (1) playful behaviour; (2) flight and distancing behaviour; (3) sociality; (4) socially threatening and fearful behaviour; (5) submissive behaviour. Using data from eight copiously represented breeds (American Staffordshire Terrier, Golden Retriever, Labrador Retriever, Lagotto Romagnolo, Nova Scotia Duck Tolling Retriever, Rhodesian Ridgeback, Spanish Water Dog, and Staffordshire Bull Terrier), they found that flight and distancing behaviours had the strongest heritability and considered their results to be potentially useful for future breeding efforts towards less fearful and more trainable dogs.
Various forms of aggression are among the most intensively investigated undesirable canine behaviours, in which, besides the influence of the environment [38], we can find ample empirical evidence for genetic predisposition (e.g., [145,146,147]). Anxiety-related problematic behaviours are also widespread in the companion dog population [148], therefore, it is not surprising that researchers try to find genetic factors that may contribute to this feature as well (e.g., [149,150]).
The increasing availability and cost efficiency of large-scale molecular genetic techniques (such as observed- and latent-trait GWAS) allow scientists to analyse vast sample sizes of genetic material. Mahmoodi et al. [151] investigated samples from more than 40 thousand dogs across more than 100 breeds, looking for candidate genes for nine behavioural traits that originated from the widely used C-BARQ instrument (e.g., [104]). These were all behaviours that are considered ‘problematic’ in companion dogs: stranger-directed aggression, owner-directed aggression, dog-directed aggression, stranger-directed fear, non-social fear, dog-directed fear, touch sensitivity, separation-related behaviour and attachment attention-seeking. The results showed that of the 41 significant SNPs identified, 26 belonged to pleiotropic neurobehavioural genes, which underlines the complexity of potential behavioural selection efforts against the listed phenotypes.
Although thus far, the research on genetically influenced behavioural traits in companion dogs has mainly focused on the problematic phenotypes, the identification of these factors obviously can help breeders to select dogs that show more favourable behavioural patterns. Directional selection towards the ‘ideal’ companion dog within the dog breeds would be a commendable effort (as suggested by some experts, e.g., [78,125]), especially because according to the results from such breeds in which the ‘working’ and ‘show/pet’ lines were recently separated, the working-type dogs usually showed more favourable behavioural dimensions than the non-working-type dogs. D’Aniello et al. [29] found that working-type Labrador Retrievers were more attentive towards humans, and Fadel et al. [44] found that working-type Labradors and Border Collies retained more of their breed-typical impulsivity characteristics than the ‘pet-line’ dogs from the same two breeds. One solution would be to identify and pursue breeding goals with the most favourable ‘companion dog features’ (e.g., [64,65]). However, in theory, there is another method that offers a faster (but still genetics-based) recipe for reaching the sought-after, ideal companion dog behaviour: combining existing behavioural phenotypes through inter-breed hybridization.
5.2. Designer Dogs for ‘Good Behaviour’? Breed Crosses
Although pedigree dog breeding by definition relies on the ‘pure-blood’ ancestry of dogs according to the studbook registries, over the last 150 years, several new dog breeds have been created from already existing breeds or landraces. Perhaps the most famous example of a breed that was created by meticulous selection and crossbreeding of contemporary landraces is the German Shepherd Dog [152,153]. The Golden Retriever is another breed that has gained long-lasting popularity among fanciers and professionals, and its origin can be found in the purposeful crossbreeding of contemporary existing dog breeds in the late 19th century, such as Newfoundlands, Irish Setters, spaniels and Bloodhounds [154]. In most cases, breeders had well-defined functional goals with these pre-planned crosses, and their intention was to achieve a ‘better’ working dog breed that combined the beneficial features of the original (parental) breeds. For example, lurchers [155] are created from crosses of herding dogs and sighthounds (e.g., Border Collies and Greyhounds); the resulting hunting dog, in theory, can be better directed due to its herding dog ancestor than the more independent sighthounds [156,157].
As each of the existing dog breeds can be characterized by a combination of breed-typical behaviours (e.g., [41,75,96]), it would also be logical if dog breeders based their efforts to create ideal companion animals on combining the favourable features of different breeds. An indirect (and not recommended) example of the potential effect of crossbreeding comes from the comparative ethology of dog–wolf hybrids. Tagliarini and Temrin [158] compared the human-directed behaviour of dog–wolf hybrids to that of Malamutes, Siberian Huskies and German Shepherd Dogs. Each research subject was kept as a companion animal in the United States, and the hybrids had 11–43% ‘wolf-content’ according to DNA-testing. While the dog–wolf hybrids showed similar playfulness compared to sled dogs, and they were no more aggressive than the dogs in the study, they were significantly more fearful than any of the other tested dogs. These results show that the behaviour of hybrids in some instances can be well predicted based on our knowledge of the parental forms: wolves are undomesticated wild animals with a long history of being intensively hunted; therefore their fearfulness towards humans [159] not surprisingly prevailed in their hybrids as well. Interestingly, fearfulness/tameness was one of the clearly segregating behavioural phenotypes in the Basenji–Cocker Spaniel crosses created in the seminal project run by Scott and Fuller [160], suggesting a rather simple genetic background for this trait in dogs. However, these authors found that other capacities in their hybrid dogs, such as problem-solving ability, were more difficult to connect with the behaviour of the parental breeds.
Deliberate crosses of established dog breeds became increasingly popular for creating sought-after companion animals towards the end of the 20th century. Often dubbed as ‘designer dogs’, they are usually created for their appearance and other morphological features rather than their behaviour [70]. The most famous cluster of designer crosses is the so-called ‘doodles’, which are hybrids of Poodles resulting from crossing with other dog breeds (for example, the Labradoodle is a Poodle–Labrador Retriever cross). Poodle crosses were invented with the original intention of making a hypoallergenic coat [161], but beyond the obvious monetary interest of the breeders who create a wide array of designer dogs, allegedly positive behavioural features are also often emphasized in the case of breed crosses [70].
Owners of designer breeds most often mention the appearance of these dogs as a driving force behind their choice [161]. Besides this, the expected positive heterosis (or hybrid vigour) effect, resulting in a generally healthier dog than the traditional purebreds, is also commonly emphasized as a reason for the increasing popularity of designer dogs. To our knowledge, this assumption has not been empirically substantiated [162,163]; on the contrary, breed crosses seem to manifest the combined array of typical illnesses of their parental breeds [164]. But what about their behaviour? Apart from a very few studies (e.g., [138]), designer crosses have not yet been evaluated for their behaviour. Unlike dog breeders living a few centuries ago, or modern behavioural geneticists who purposefully target functional (working) performance [143], modern designer dog breeders apparently do not specifically aim to change the behaviour produced in these crosses relative to that seen in their originating breeds. Based on Mendelian genetics, one could assume that F1 breed crosses will exhibit ‘intermediate’ behavioural phenotypes of the parental breeds. Unfortunately, no such studies have been conducted so far on this topic comparable to the classic projects Scott and Fuller did [160], and the investigation by Shouldice et al. [165] showed no relevant difference between the C-BARQ behavioural profile [166] of Goldendoodles, Labradoodles and their parental breeds (Labrador or Golden Retrievers and Poodles). From the aspect of ‘ideal companion dog behaviour’, designer crosses that are created with thoughtful planning can have great potential value; however, a deeper understanding of the inheritance of those behaviours that play a key role (either positive or negative) in dog–human interactions would be of paramount importance.
5.3. Looking into the Crystal Ball of Behavioural Ontogeny: Puppy Tests
Dogs spend most of their lives as adult animals with their owners; therefore the match between the dog and the expectations/needs of humans is predominantly evaluated based on the behaviour of adult dogs. Since the seminal research of Scott and Fuller [160], the recommended time window for acquiring a puppy has been believed to be around the eighth week of life. However, because of more recent considerations suggesting that young dogs could greatly benefit from the social environment of their mother and littermates [28,167], nowadays there is a tendency to place puppies with their new owners no sooner than 10–12 weeks of age (e.g., [65]). At the same time, regardless of the future role of the dog (i.e., working dog or companion), it would be highly beneficial if: (1) based on the puppies’ behavioural traits, the ‘perfect match’ between the prospective owners’ needs and the dog could be better guaranteed; and (2) the upbringing of puppies could safeguard against problematic behaviours and generally unsound behavioural traits later in life. From this perspective, professionals and researchers have proposed and tested several methods for either improving the stimulative environment of young dogs while they are still with their mother [168] or evaluating puppies’ behaviour to forecast their suitability as adults for specific purposes/needs (e.g., [169]).
Puppy tests could help predict subsequent (adult) behavioural characteristics in at least two ways. The seemingly simpler expectation is that, based on the puppy’s temperament/behaviour, it would be possible to assign ‘the right dog to the right owner’ (e.g., [170]). These requests often go no deeper regarding dog behaviour than trying to forecast whether some puppies will become highly active/high drive, or rather laid-back specimens of their breed. According to studies, in such cases, unexperienced prospective owners would greatly benefit from simply visiting the litter of puppies with their mother present to get a glimpse of the expected behaviour of that breed and breeding line (e.g., [39,171]). However, sometimes it would be desirable to foresee much more specific information by performing behavioural tests on the puppies, such as whether they would become suitable for tasks that some of the armed forces require or to become successful candidates for sophisticated assistance-dog training [172,173,174]. Junttila and colleagues [175] found that some of their testing battery tasks were suitably predictive of adult dogs’ corresponding behavioural traits when they tested young dogs between 3 and 7 months of age. Following human pointing gestures predicted adult trainability, while performing well in a detour task predicted energy level and impulsivity, and performance in the ‘unsolvable task’ was associated with adult fear of strangers. Although in that study the young subjects had already passed the typical ‘puppy age’ when most dogs enter their new owners’ home, it seems reasonable that with properly chosen temperament and personality tests, one could find behavioural traits that stay relatively stable from juvenile age to adulthood (e.g., [176]). Unfortunately, in reality, these expectations cannot often be justified (e.g., [177,178]).
One reason for these disappointing results could be that the behavioural repertoire and typical responses of puppies can be vastly different from the adult working/sports dog typical task-related behaviours that we wish to predict (e.g., [172]). As puppies usually spend only a relatively short time at the facility where they were born, and their breeders and future owners both wish to avoid misplacement of young dogs who will later turn out to be unsuitable for the desired role, the ‘ideal’ puppy test should be feasible to perform as early as possible [170,177]. This increases the difficulty of this kind of procedure, because the younger the dog is, the simpler the behavioural reactions it can be tested for are, which do not necessarily resemble the adult behavioural pattern that needs to be predicted [172]. Thus, it is no surprise that the currently highly sought-after ‘Holy Grail’ of puppy tests (‘early—simple—highly predictive’) to the best of our knowledge, has not yet been discovered and published.
6. Dog Breeds in the Future: A Proposal and Evaluation of Four Potential Avenues for Their Development
In the previous chapters, we outlined that the age of modern (post-Victorian) dog breeding created not only a bewildering array of dog breeds, but towards the end of the 20th century it also helped to develop a never-before-seen popularity of companion dog keeping all around the world [102,179]. The rise of urban dog keeping has resulted in a paradoxical situation: people can now choose from hundreds of established dog breeds, which have a wide enough morphological and behavioural variability to seemingly please everyone—still, modern companion dog keeping is plagued by a fast-growing occurrence of behaviourally problematic dogs (e.g., [180,181]). The mismatch between owners’ expectations, capabilities, and the acquired dog’s temperament/behaviour likely leads to dissatisfaction and negative experiences for the owner, relinquishment of the animal, and consequently compromised human and animal welfare [181]. We have seen that in theory, there are several solutions at hand, and these could help prospective dog owners to make the best choice and know what to expect from their new canine companion. Scientific research has highlighted several behavioural and cognitive characteristics of dog breeds based on their ancestry and functional selection (for a review, see [41]), which can provide basic guidelines during the planning stage, regarding the behaviour of these breeds. Breed crosses could have some potential for pre-planning based on the behaviour of the parental breeds. In the case of separate working and pet/show/sport lines of the same dog breeds, some scientific evidence is also available to make an educated guess about their expected behaviour as well [44,182]. We should also mention the indispensable role and importance of the wide array of training methods, many of which have been found to be effective for improving the ease and quality of amicable dog–human interactions (e.g., [168,183,184]). However, these methods and information sources have been available for quite some time, without any remarkable advancement in the quality of dog–human relationships from the aspect of approaching an ‘ideal dog’.
The controversies and problems surrounding purebred dogs’ health and compromised welfare due to their extreme anatomical features have already resulted in some pioneering publications with potential solutions for future breeding efforts (e.g., [70,125]). In theory, the fundamental attributes of a highly successful urban companion dog have also been collated by several recent surveys [64,65,185]. Accordingly, dog owners would prefer an ‘amicable’ dog that is safe with children, does not show aggression towards humans and other animals, can be left at home without problems for longer periods of time, does not bark excessively, and has a laid-back temperament. Although some dog breeds’ behavioural characteristics may fall closer to this idealistic canine companion than others, we are not aware of a scientific study or applied program (except those for working dogs, e.g., [91]) that has attempted to get closer to an ‘improved’ companion dog. The public need for a scientifically grounded, large-scale program for selecting and breeding dogs with equally sound health and behavioural traits is further highlighted by such initiatives as “The Functional Dog Collaborative” https://functionalbreeding.org/about-us/ (accessed on 22 August 2026) led by distinguished researchers and canine professionals. Urged by the high occurrence and worrisome outcome of problematic (mismatched) owner–dog relationships [181] and based on the apparent inadequacy of our current possibilities towards the improvement of this situation, we believe that our societies should move towards a new era of science-based functional dog breeding. In this chapter we outline several potential avenues for developing such dogs (‘breeds’) that equally satisfy the various end users (generally: companion dog owners) and provide good welfare and safe lives for the dogs themselves. We are also going to evaluate each of the proposed strategies with the help of the SWOT (Strengths, Weaknesses, Opportunities, Threats) principle, outlined for applied ethological studies by Van Belle et al. [186]. This analytic tool provides a straightforward framework for a multiple-aspect assessment of potentially ‘competing’ solutions (e.g., [41]), thus helping to recommend considerations for future research and decisions.
The visual illustration of the four proposed strategies can be seen in Figure 1.
[图：Schematic illustration of the four proposed breeding strategies to achieve predictably behaving dog breeds that are suit]
6.1. Strategy ‘Converge’: Converging Most Dog Breeds Towards the ‘Amicable Urban Dog’
This avenue has already been suggested by those authors who surveyed the Australian and Italian public regarding their preferences for behaviour in a canine companion [64,65,185]. Once the goal is set, in theory, breeders can exert strong directional selection on basically any breed to reach as close a match to the desired behavioural phenotypes as possible. One can assume that there will be such breeds that right from the start fall closer to some of the chosen milestones of an ideal urban dog. Others, especially such breeds that were selected/used for guarding, personal protection, fighting with other dogs/animals, or show a high drive for either cooperative (e.g., herding dogs) or individual work (e.g., terriers, Dachshunds), would initially fall short of many requirements. We should also keep in mind that beyond the results of behavioural assessments, opinions concerning a breed’s suitability for the role of being a friendly family companion can be strongly shaped by breed-related stereotypes and ad hoc encounters with positive or negative experiences [119].
Most research efforts were aimed at identifying dog breeds with the highest and lowest levels of robust phenotypes such as fearfulness, aggression, separation-related problems and excessive levels of activity (e.g., [75,187]). McGreevy et al. [112] found that in general, the smaller-bodied a dog breed was, the more frequent and intense signs of behavioural problems it showed (including fear and aggressive reactions, or problems with being separated from the owner). A Polish study highlighted that while larger-bodied dogs were more commonly reported to have problems with aggression, smaller dog breeds were reportedly more ‘hyperactive’ than others [188]. Dogs that were selected for working in close cooperation with their handler showed a stronger inclination to express signs of separation problems than the independent working breeds [132]. Ancient dog breeds (such as the Malamute, Siberian Husky and the Akita) can show strong inclinations for being aggressive with other animals, too active, and disturbingly vocal [189]. Thus, in general, each dog breed or breed type can show undesirable features from the aspect of the ideal companion dog. According to the strategy ‘Converge’ towards developing the ideal companion dog for the future, basically every dog breed requires moderate-to-strong directional selection to achieve this goal.
The main ‘philosophy’ behind this proposed avenue towards future dog breeds is that it basically puts almost all the responsibility on dog breeders and indirectly on the dogs themselves. In theory, once the required ‘user-friendly’ dogs have been created, prospective dog owners would be safe to acquire them without concern for the breed’s previously expected behaviour.
Strength: If a large proportion of dog breeds were to be selected to be an ‘ideal companion dog’, this would significantly elevate the general safety of ‘users’ (dog owners and their families) as well as their social environment (other people and animals). We can expect this based on the notion that human- and animal-directed aggression is one of the most serious problematic behaviours in dogs often with dire consequences (e.g., [190,191]). The widely available ‘amicable’ dogs from a large selection of breeds would positively affect the experience of dog keeping, because problematic dogs cause considerable emotional and monetary hardships in the human environment [181]. We can also expect that a more stress-free dog–human coexistence and lower occurrence of such problem behaviours such as separation-related issues would elevate the welfare of dogs as well [192]. Finally, the strength of this strategy is that we have already collected a considerable wealth of empirical knowledge about the various aspects of dog breed behaviour, which can provide a solid starting point for the outlined plans.
Weakness: In theory, each dog breed can have genetic flexibility that would allow directional selection towards the requirements of an amicable companion dog, and we also should not forget the positive effect of additional training by the owners as a further approximation of the desired companion dog (e.g., [33]). At the same time, large-scale behavioural comparisons among convenience sample dog breeds (e.g., [193]), as well as breed-type comparisons based either on functional breed selection [194], or breed group ancestry (e.g., [189]), showed that there are strong enough differences among the typical behaviours and socio-cognitive capacities of dog breeds that would probably make it a near-futile endeavour to select ideal companion dogs from particular breeds or breed types. Especially today, those unwanted characteristics that were once considered most useful for a given breed’s function seem difficult to ‘tone down’ by selection. For example, heightened vigilance, vocalization and territorial behaviour of guarding breeds, or the ‘hyperactivity’ of terriers, are such trademarks of the original work-directed selection that would be suboptimal in an ideal urban companion dog. Most likely, only a proportion of today’s dog breeds could provide an easy starting point for further selection towards being an ideal urban companion. As the main target traits (aggressivity/fearfulness, hyperactivity and proneness to separation-related problems) can be connected to somewhat different genetic backgrounds (e.g., [147]), the selection against each of them at the same time might be very difficult. Therefore, breeders would likely have to choose which feature of their breed would be the most important to improve first. The randomly bred mixed-breed (‘mongrel’) dogs represent another weakness in the framework of the ‘Converge’ strategy. Mixed breeds comprise a large proportion of companion dogs everywhere (for example, in the United States, nearly 50% of the dog population was reportedly non-purebred [195]). In the case of mixed breeds, we do not have extensive a priori knowledge of the dogs’ expected behavioural traits; thus using them for effective directional selection of ideal companion dogs may represent a prohibitively difficult task.
Opportunity: The strategy ‘Converge’ may offer a ‘last chance’ for such breeds that became unpopular or nearly extinct because of their original work function, and thus some of their behavioural characteristics turned out to be incompatible with urban companion dog keeping. Large livestock guardian breeds (LGDs) which were recently re-introduced to perform their original task of flock protection against threats posed by wolves, coyotes and golden jackals (e.g., [196]), would otherwise be difficult to integrate into companion dog status. These breeds can easily harm people due to their large size and past selection (e.g., [197,198]), and partly because of this, they are losing popularity as companions. The ‘Converge’ strategy could encourage dog breeders to develop a more compatible version of LGDs for companion purposes. The same opportunity would be true for other dog breeds with a strong propensity for being protective (e.g., Rottweiler, Boxer, German Shepherd Dog, Dobermann, etc.)—although these are not threatened by extinction, they can be problematic due to reports of their human-directed aggression (e.g., [199,200]). The opportunities of the ‘Converge’ strategy would be aided by the already ongoing processes of disruptive selection in particular dog breeds, where distinct ‘working’ and ‘show’ lines exist [44]. These lines can reveal the behavioural variability within the breeds and readily offer more raw material for further selection towards ideal canine companions. From the perspective of prospective dog owners, the ‘Converge’ strategy could also provide a favourable outcome for those that prefer such dog breeds that were originally unpopular in the community due to their negative reputation (e.g., [201]). Finally, as the strategy ‘Converge’ necessarily includes the development and application of appropriate behavioural testing, these tests could also be used for mixed-breed dogs, thus providing insight for personnel at shelters, rescue organisations, and prospective owners regarding the suitability of these dogs as urban companions.
Threat: The ‘Converge’ strategy indirectly poses a threat towards dog breeds that fail to show the desirable improvement towards becoming ideal companion dogs. If prospective owners generally expect predictably ‘amicable’ behaviour from the available breeds, those breeds will face a rapid loss of popularity in which the breeders are reluctant to select family-friendly lines, or the lack of the required genetic background undermines the efforts of the breeders. Such dogs would also face hardships in the case of adoption from shelters (e.g., [202]). A still-unknown magnitude of threat in the ‘Converge’ strategy would be the unpredictable loss of some breed-typical behavioural traits due to the directional selection towards companion dogs. There are indications that the ‘show lines’ of some breeds, which lacked the original work-related behavioural selection in the last decades, exhibit more similar behavioural traits to each other than their working-line counterparts [44]. Once these functional traits are lost, we may face a situation where the resulting dogs will also lose a significant part of their previous appeal. In the case of a widely successful ‘Converge’ strategy, users of original (functional) working dogs could face difficulties finding a wide enough pool of this type of their favourite breed. Another potential threat of a different nature could be the largely unexpected emergence of potentially unwanted phenotypes that have a genetic, neural or physiological connection with some of the favoured characteristics of ‘amicability’. A functional analogy is provided by research on silver foxes that have been selected for tameness only; however this directional selection also affected their morphology and non-targeted behaviours, such as their vocalizations [203,204].
6.2. Strategy ‘Diverge’: Lessening the Breeds’ Behavioural Unpredictability by Strengthening Their Breed-Typical Traits
Most behavioural phenotypes of the dog are strongly influenced by the environment (i.e., early socialization [28], training [205], and previous trauma [206] etc.), and epigenetics (e.g., [207]). However, recent molecular genetic advances (e.g., [208]) still suggest a detectable genetic background for signature behavioural characteristics of the main working types of dogs (such as herding, predation and trainability). The core idea behind the strategy ‘Diverge’ is not different from the generally recommended ‘best practice’ for today’s prospective dog owners: after careful examination of the possibilities and learning the existing characteristics of a dog breed, only then make an educated choice from the best-matching breeds [80]. Current research shows that dog owners indeed use a wide variety of sources of information about both the characteristics of dogs and requirements for owners (e.g., in the UK: [209]). However, as Mead and colleagues [209] noted, this information about breeds is often controversial, and according to another study, “choosing dogs on a whim” was still one of the most often reported methods used by owners [82]. If we add to this the often strongly varying behavioural profiles of dog breeds (e.g., [210]), further strengthened by the divergence due to breeding line-based disruptive selection [44], the rationale behind the ‘Diverge’ strategy becomes obvious: if dog breeds showed a more ‘clear and consistent’ behavioural profile, this predictability could positively reward those dog owners who otherwise paid keen attention to matching their needs and possibilities with the chosen breed’s characteristics. It is worth noting that the strategy ‘Diverge’ uses different breeding techniques (stabilizing selection vs. directional selection) than the ‘Converge’ strategy did and also emphasizes the main responsibilities differently. While the ‘Converge’ strategy put the main responsibility onto the breeders, the ‘Diverge’ strategy highlights the responsibility of the prospective dog owner and offers help to them by clarifying the (predictable) differences between the various dog breeds’ and breed type’ behavioural traits.
Strength: Perhaps the ‘Diverge’ strategy can build on the widest base of existing empirical knowledge about the various dog breeds’ behavioural characteristics. There is a rich variety of research and review papers that approached the breed-related behavioural differences from the aspect of molecular genetics (e.g., [71]), functional selection [41], or genetic distance-based ancestry [211]. Based on the practical descriptions of the original work tasks and the behavioural traits of dog breeds for fulfilling these requirements, aided by the toolkit of applied ethology in behavioural evaluation (e.g., [25,68,193]), breeders could select the most ‘breed-appropriately’ behaving subjects for further breeding. The same ‘breed-stereotypical’ knowledge could provide guidance for prospective dog owners, who would acquire such dogs from breeds that should fit their expectations.
Weakness: One of the main problems with the ‘Diverge’ strategy is that many first-time and experienced dog owners have inadequate knowledge about both responsible dog ownership [212,213] and the needs and characteristics of the given dog (breed). Studies show that when choosing their next dog, owners of extreme brachycephalic breeds would opt again for the same breed even though they are aware of the compromised health and welfare of these animals [133]. The attitude of such owners who are willing to accept and normalize, for example, ‘laziness’ as a breed-specific characteristic of Pugs or bulldogs, instead of regarding it as a serious consequence of cardiac and respiratory malfunctioning [214], can be a widespread phenomenon for other problematic but commonly ‘accepted’ breed-related behavioural and structural anomalies as well. Thus, the success of the ‘Diverge’ strategy mainly depends on the efficiency of owner education about what the desirable and typically expected behaviours are in each breed. A further weakness in the ‘Diverge’ strategy is the extensive variability in many breeds’ behaviour—which is further exaggerated by the rapidly diverging ‘working’ and ‘show’ lines. For example, Overall et al. [215] found that noise sensitivity showed not only breed-, but also family-line differences among Australian Shepherds, Border Collies and German Shepherd Dogs. Strong within-breed behavioural variability warrants the need for extreme caution when the ‘true’ breed characteristics are being agreed upon by breeding organizations. Finally, there is a further weakness in the strategy ‘Diverge’ as it lacks guidelines for establishing the desired, typical behavioural profile for mixed-breed dogs.
Opportunity: Compared to the ‘Converge’ strategy, here theoretically, each breed could have a future in its original functional form—given that society and the ‘end-users’ still need them. It is likely that most breeds can be adapted to the companion dog role if the owners receive reliable and useful guidance for their needs and characteristics, and they are also educated about the realistic assessments of their own possibilities and responsibilities as an owner [212]. The key to the success of the ‘Diverge’ strategy is making the right choices by prospective dog owners, and for these choices to impact the breeders’ efforts towards providing reliably and predictably behaving dogs according to the breed standards. If breeders who follow the recommendations of the ‘Diverge’ strategy succeed in stabilizing the breeds’ characteristic behavioural traits, it would not only be beneficial for prospective companion dog owners for making educated choices, but also the users of working dogs could benefit from it as well [27,143]. Although the genetic predisposition (heritability) of many behavioural traits is still poorly understood, and the role of ontogeny and environmental factors should not be underestimated [27], breeders and trainers of working dogs would probably have an easier task if they could start with breeds that have more clearly defined behavioural profiles. Regarding the mixed-breed dogs, as the ‘Diverge’ strategy does not specify the ubiquitous behavioural blueprint for the ideal companion dog, the wide and unpredictable variability of mongrels could also be utilized after proper behavioural profiling. Those prospective owners who prefer adoption of mixed breeds could also find better-matched companions if there was more reliable information about their main behavioural traits. Archer et al. [216] showed that the adoptability of mixed-breed dogs from shelters was more positively affected by providing temperament/behaviour profile descriptions along with pictures of the dog than providing the same photos with tentative breed labels.
Threat: A potential, worrisome outcome of the strategy ‘Diverge’ is the further increase in the level of inbreeding across breeds, if breeders use stabilizing selection for the behavioural traits without effective measures of preserving the genetic variability of the dogs (e.g., [217,218]). Instead of aiming towards a generic amicable behavioural profile, the strategy ‘Diverge’ emphasizes the strengthening of original, function-appropriate behaviours of the breeds. Thus, those breeds whose original employments have become obsolete, and whose typical behaviour is mainly unsuitable for companion animal purposes, would be more likely to disappear. Probably the most relevant weakness of the ‘Diverge’ strategy is the lack of a well-defined ‘ideal urban dog’, because this strategy relies on the (idealistic) image of educated and well-prepared prospective dog owners, who will choose from among the predictable and reliably behaving breeds correctly. Considering the current disappointing tendencies of highly popular dog breeds to also be aligned with a long list of morphology-, genetics-, and behaviour-based health, welfare and safety problems [101,102], it would probably be a naïve expectation to believe that dog owners would opt for such breeds that fit best their and their environments’ possibilities. Another weakness of the ‘Diverge’ strategy is that if the education of the owners and breeders is misguided or less effective, the inappropriate emphasizing of breed-specific features can have an unwanted effect of further popularizing/normalizing both the behavioural and morphological extremities of some breeds.
6.3. Strategy ‘Split’: The Creation of Well-Distinguishable Working- and Urban Companion Variants Within the Same Breed
If we consider the main threats of the ‘Converge’ and ‘Diverge’ strategies, it becomes clear that while the former indirectly implies the eventual loss of originally prized working dog behavioural traits, the latter likely fails to provide a wide selection of ideal city-compatible companion dog breeds. The ‘Split’ strategy can be regarded as a combination of the previous two proposed methods, with a dual aim for how to improve the behaviour of the targeted breeds. On the one hand, the goal would be to develop a reliable ‘amicable’ type; on the other hand, a reliable working line would be either restored or maintained, depending on the actual behavioural status of the existing breeding stock.
A good model for such dog breeds where the ‘Split’ strategy could provide a generally useful solution can be found in the large livestock guardian breeds (LGDs). According to recent molecular genetics results, these breeds (originally as landraces) have been developed across Europe and Asia [219] to fulfil basically the same purpose: to protect various livestock from any dangerous predators and from theft [220]. These dogs show extreme loyalty and no predatory behaviour towards livestock [135], but understandably, their aforementioned work-related behavioural traits (and large size) hardly suit urban companion status. When and where large predators (such as wolves) were hunted to extinction, LGDs lost their original jobs and their numbers, along with the knowledge of their management, were also considerably reduced (e.g., [221]). When the natural predator species started to return to their original habitats, due to the applied biological conservation programs, the need for reliable LGDs opened up again [196]. Currently, LGDs are considered not only one of the most effective but at the same time the most economical and wildlife-friendly solutions for extensive livestock management [222]. However, nowadays probably more LGDs are bred for reasons of fancy (kept for showing and/or as a companion) than for protecting livestock. Most LGD breeds are considered to be a less suitable choice for being a family companion (e.g., [223]), although their impressive appearance and size appeal to many [82]. In theory, for dog breeds such as the LGDs, where there is still a need for fulfilling the original role, which in turn is not fully compatible with the companion status, the ‘Split’ strategy could provide a satisfying solution for both types of utilization.
Strength: The main advantage of the ‘Split’ method is that it can be basically used on any dog breed with larger flexibility than the ‘Converge’ or ‘Diverge’ strategies. An initial population-wide assessment is needed to clarify: (1) whether there is a need for retaining the original function-specific behaviours of a breed; and (2) if this function is significantly different from the requirements of an ideal companion dog’s behavioural repertoire. The dog breeds that are undergoing the ‘Split’ method can differ in which breeding goal (‘companion’ vs. ‘original work’) requires more breeding effort to reach. Adopters of the ‘Split’ strategy can rely on the already existing breeding lines in some dog breeds (‘show’ or ‘working’ lines, e.g., in the Kelpie: [224]; in Border Collies and Labrador Retrievers: [44]).
Weakness: Presently it is unknown whether in some dog breeds that have already been strongly selected for a specific task/temperament, there is enough genetic variability left to reach considerable results by breeding them towards a totally different functional role. Many dog breeds suffer from a serious loss of genetic diversity (e.g., Leonberger: [225], Dobermann: [226]), which indicates that it can be problematic to find a sustainably wide genetic basis for breeding dogs for more than one behavioural type in each breed. Another general weakness of this strategy is that it does not offer a place for mixed breed dogs in the breeding or evaluation programs.
Opportunity: Compared to the ‘Converge’ and ‘Diverge’ methods, the ‘Split’ strategy would probably meet with the widest and easiest acceptance among the enthusiasts of the various dog breeds. Instead of moving the entire breed towards a predefined, uniform behavioural goal, with the ‘Split’ strategy it would be theoretically possible to establish the ‘companion’ and ‘working dog’ versions of the same breed. This would not only provide suitable dogs for a wider array of owners, but it would also help to maintain higher genetic diversity throughout the whole breed.
Threat: As the ‘Split’ strategy is intended to eliminate the problems that could arise from the adoption of either the ‘Converge’ or ‘Diverge’ methods, it has far fewer risks than those options. By allowing and encouraging the breeding of both the original breed purpose dogs and developing the new, amicable companion type, there is no threat of losing the traditional work functions or not having an urban-compatible version of the given breed.
6.4. Strategy ‘Merge’: The Creation of New Urban Dog Breeds
This avenue can be considered to merge the best possible outcomes of some of the previous strategies while also allowing the breeders and dog owners to opt for more traditional solutions. The main problem with the strategy ‘Converge’ was that both breeders and breed fanciers could be reluctant to accept changing their favourite breeds to a uniformly ‘amicable’ urban dog. In the case of strategy ‘Diverge’, as its success relies on the owners’ educated choices from among dog breeds that are not necessarily suitable for urban companion status, there is a good chance that they would make suboptimal decisions. The strategy ‘Merge’ emphasizes the need for newly created breeds, as the main goal of selection would be to match the requirements of an easy-going, amicable urban companion dog, most likely available in a few different sizes and appearances (e.g., coat type, ear position, etc.). By creating new breeds, fanciers of the traditional breeds would not feel that the characteristics of their favourite dogs would be threatened, and those prospective dog owners who simply need an ideal urban companion without excessive preparation and the need for training, could also find a straightforward solution.
We must admit that among the four strategies we have outlined, ‘Merge’ would probably be not only the most promising but also the most challenging. Those dog breeds that were created through planned crosses and careful subsequent selection in the first half of the 20th century were mainly built according to specific functional needs as working dogs (e.g., German Shepherd Dog, Moscow Watchdog, Czechoslovakian Wolfdog). Currently, crossbreeding between some of the popular working dog breeds sometimes occurs with the intention of maintaining robust health through higher genetic diversity (e.g., Labrador and Golden Retrievers in service and guide dog programs: [227]). So far, to the best of our knowledge, no dog breed has been selected with the goal of becoming the ideal companion dog. The recent popularity of the exponentially growing inventory of ‘designer crosses’ shows more the ingenuity of their creators as marketers to meet buyers’ preference for unusual looks than any planning for positive behavioural traits [70]. Thus, those breeders who would embark on the task of establishing new breeds based on their amicable features would need to (1) find the most promising ‘starter’ breeds (similar to the ‘Converge’ strategy); then (2) create breed crosses, and with directional selection, move towards the best-balanced features; and finally (3) stabilize the characteristics of the dogs. Only after the new ‘breed’ became reproducible with reasonably stable behaviour and appearance could its popularization among prospective owners begin.
Strength: With this strategy, owners could access functionally bred dogs, which would reliably show the behavioural traits that are ideal for the urban companion role. These new types would offer favourable alternatives to the traditionally preferred companion breeds, with the advantage of having multiple ancestor breeds, which may result in higher genetic diversity [228]. The high number of currently existing dog breeds could offer wide and sufficiently variable raw material for the initial selection, especially if it were to be aided by molecular genetic results identifying the most-sought (and unwanted) traits. There are already promising results regarding the heritability and genetic variability of human-directed social behaviours (e.g., [229,230]). Another advantage of the ‘Merge’ strategy is that while it could add new breeds to the market, it would leave the existing dog breeds in their current behavioural diversity (similar to the ‘Diverge’ strategy); thus, it would not raise strong resistance from breed enthusiasts who would not like to see the alteration of their breeds’ original characteristics.
Weakness: This strategy would probably be the most difficult to establish on a larger scale. Breeders of the new ‘amicable’ breeds could face the problem of registering their dogs, as the main kennel clubs do not embrace such endeavours. While the ‘Converge’ and ‘Diverge’ strategies did not involve crossbreeding, thus the resulting dogs could be easily registered as traditional purebreds, the success of the ‘Merge’ strategy would be highly dependent on the enthusiasm, marketing and education of both breeders and dog owners. Designer crosses are not unequivocally welcomed among dog breed fanciers and breeding organizations (e.g., [164]), and this is a warning sign for the purpose-bred new ‘amicable’ breeds, as well. A further weakness of the ‘Merge’ strategy would be its slow pace. If breeders were to aim to create ‘stabilized’ new companion breeds, it could take several generations of meticulous selection (with an average of 3 years per generation) until the behavioural traits could be stabilized. As a comparison, it took no less than 25 years to stabilize the Czechoslovakian Wolfdog after the initial and subsequent crossings with wolves [231]. Even with an optimal start and widespread education/marketing programs, we can expect that the strategy ‘Merge’ would remain a small-scale ‘connoisseur’ option for the most devoted enthusiasts for a long time. This would leave the current problems of choosing less suitable dogs as urban companions mainly intact.
Opportunity: By offering brand-new dog breeds that are characterized by the most suitable amicable companion dog traits, the strategy ‘Merge’ has the advantage of gaining an enthusiastic supporter base more easily than the other three strategies. The newly developed breeds would not give the impression that the selection towards urban companions would endanger the original qualities of the traditional breeds. The opportunity to educate dog fanciers about the desirable features of an urban companion dog, as well as propagating responsible dog ownership, could gain new momentum if it were connected to the introduction of the newly selected breeds. It would be easier to convince dog fanciers of the benefits of amicable companion dog behaviour when this is associated with a new type of dog, rather than argue that the long-established dog breeds should undergo fundamental behavioural changes. It would be possible to use mixed-breed dogs in the breeding program of new amicable dog types, because through behavioural testing and genetic assessments, those dogs that have some of the desired qualities could also become valuable breeding sources.
Threat: This strategy, even if it is successful, would provide small-scale access to the newly developed companion dogs only after a long time. The demand for the sought-after ideal amicable dogs could create an opportunity for unscrupulous breeders to offer low-quality, poorly bred dogs of questionable origin, similar to the tendencies in the breeding of today’s fashionable breeds [128]. This phenomenon would also be intensified when the positive experiences, and successful publicity about the new amicable companion breeds result in high purchase prices. We should also not forget the conservative attitude of many dog fanciers, making them reluctant to change their dog breed preferences. This is well demonstrated by the fact that despite multiple empirical results showing the complex compromised health and welfare status of extreme brachycephalic breeds (e.g., [232]), their devoted owner population resists ceasing to choose them [214]. For different reasons, supporters of adopting troubled (and mainly mixed-breed) dogs from shelters would probably also resist opting for the new companion dog breeds.
Table 1 shows the main details of the SWOT assessment for each strategy.
Table 1. The summary of the SWOT analysis in the case of each strategy proposed in this article.
| Converge | Diverge | Split | Merge | |
|---|---|---|---|---|
| Strength | As a result, widely available amicable urban dogs from many breeds. | The widest involvement of breeds and existing breed-related knowledge. | Flexible application to most dog breeds. Already existing within-breed lines. | Offering new amicable urban dogs and leaving the original behavioural traits intact in the ancestral breeds. |
| Weakness | Selection for ‘amicability’ would not be feasible in many breeds. Parallel selection for several preferred traits would be difficult. | Extensive need for educating the prospective owners. Often strong within-breed behavioural variability. | In the case of already low genetic variability, the dual selection purpose might not be feasible. | Slow progress is due to the difficulty of creating new ‘breeds’ with a targeted behavioural complex. Strong resistance from traditional breeders’ organizations. Possibly small-scale output. |
| Opportunity | ‘Last chance’ to save some rare or less popular breeds. Developing such behaviour tests that can also be used for assessing shelter dogs. | Widespread availability of reliable working dogs. With the right education, a safe choice for companion dog owners. | The highest acceptance among the various enthusiast groups. Good chance for retaining maximum genetic variability within the whole population. | The de novo created urban dog breeds would be suitable for effective education about the advantages of amicable dogs. Less friction with traditional breed enthusiasts. Easier involvement of mixed breeds for the breeding program. |
| Threat | Random loss of such, often-preferred behavioural phenotypes that characterized the original working dogs. Unpredictable emergence of unwanted new behavioural traits. | Disappearance of obsolete function breeds. Strong chance of loss of genetic diversity. Purpose-bred amicable companion dogs will be missing. | Not applicable, as the ‘Split’ strategy is intended to attenuate the threats that may affect the ‘Converge’ and ‘Diverge’ strategies. | High demand for the available new amicable dogs, and hence opportunistic suboptimal breeders’ practices. Ethical issues with potentially compromising the adoption of shelter dogs. |
| Main steps | Identify the most suitable breeds. Repeated rounds of directional selection and behavioural assessment. | Identify optimal breed functions. Repeated rounds of stabilizing selection and behavioural assessment. | Assessment of within-breed behavioural variability. Depending on the prevalence of ‘amicable’ and ‘working dog’ traits, applying disruptive selection to achieve both goals. | Behavioural and, ideally, genetic assessment of initial breeding animals to identify the most suitable dogs. Repeated rounds of directional selection and behavioural assessment. |
7. Discussion
Companion dogs enjoy a never-before-seen and still steadily increasing popularity around the world [53]. Besides the undeniable and mutual benefits of human–dog interactions [233], the magnitude of problems arising on both the animal and human side is also alarming [181,234]. Among the many potential influencing factors, the ever-growing proportion of urban dog keeping plays a crucial role. Many dogs seem unsuited to urban housing conditions and poorly accommodate the busy lifestyle of their owners, which comes with long periods of isolation (e.g., [64]). Parallel with this, as dog ownership became a societal norm or almost a fashion (e.g., [235]), the average knowledge base of dog owners regarding the needs of dogs and their own responsibilities remains inadequate [236]. In this opinion paper, we present a thought-provoking solution: could we amend the above-highlighted problems by placing emphasis on breeding such dogs that are either (1) more suited to the criteria of the ‘amicable urban companion’ (by following the Strategies ‘Converge’, ‘Split’ or ‘Merge’); or (2) acquire a more reliable and predictable behavioural profile according to their original breed function (Strategies ‘Diverge’ and ‘Split’).
These strategies are aimed at the modification of behavioural traits of dogs by mainly using the assets of artificial selection. Obviously, various training methods can offer faster and more readily available tools for influencing the behaviour of most dogs (e.g., [237]). We do not claim that by choosing a dog or dog breed that is the best possible match to the expectations and possibilities of the owner, this could fully substitute for the subsequent need for training and joint activities with the dog. However, there is empirical evidence that various dog breeds respond to human teaching efforts with different intensity (e.g., [43]) and efficacy (e.g., [42,194]). Dog breeds that were selected for performing cooperative tasks with humans learned more keenly and better from a human instructor [42], and they followed human communicative signals more effectively [79]; but they were also more prone to separation-related problems [132]. Furthermore, not only functional selection, but the genetic distance from the dogs’ common ancestor can also influence the dogs’ responsiveness to human presence and communication. According to a questionnaire study that used the C-BARQ instrument, primitive (ancient) dog breeds were found to be less attentive and showed weaker attachment towards their owners [238]; and based on social learning performance from a human demonstrator, modern gundogs (e.g., retrievers, pointers, spaniels) and utility breeds (German Shepherd Dog, Hovawart) were the most attentive and reliant on human behaviour [25,239]. Thus, we can conclude that currently, the various clusters of dog breeds, along with the enormous variability of mixed-breed dogs, can make it very difficult to outline a generally reliable and easy-to-use training regime for those who wish to adapt these dogs to the requirements of being an ‘ideal urban companion’. Therefore, we believe that the initiative of creating more fitting dogs (breeds) via purposeful behavioural selection would be a useful starting point for having fewer problems and more enjoyable coexistence with our companion dogs in our cities.
Clearly, none of our proposed strategies could work without the widespread, coordinated efforts of dog breeders, and preferably, international breeding organizations. Such initiatives have been suggested for supplementing high-quality detection dogs [91]; to improve canine health, well-being and welfare [240], and to solve the problems of the genetic health of purebred dogs [241]. However, regarding the organized efforts to improve the behaviour of dogs to meet public demand, we only know of a few, relatively small-scale interventions. One of these was documented in the Netherlands, where Rottweilers needed to be bred for low levels of aggression and fear to avoid legislative banning. According to the reports, when the registering of Rottweiler litters was only available if the parent dogs successfully passed the official behaviour tests, eventually the Dutch Rottweilers started to show improvements in their aggression and fear indices [200]. Breeding for a broader set of behavioural traits, such as the requirements of an ‘amicable companion dog’, would not only need clear, consensus-based target behaviours and suitable testing methods for the assessment, but also many devoted breeders aligned with a coordination council that would provide the necessary guidance and promote cooperation among the breeders. This would be especially true for the strategies ‘Converge’ and ‘Merge’. The latter is aimed at developing new amicable urban dog breeds, while the former encourages the participating breeders to perform directional selection within the existing dog breeds towards amicable traits. In these cases, success would depend on the proper and standardized behavioural evaluation of the new generations of dogs created. The strategies ‘Diverge’ and ‘Split’, would be easier, because every involved breed (or a sub-population of the breeds in the case of ‘Split’) would need to be stabilized around the typical, functional behavioural traits described in its standard and supported by practical evidence.
The Novosibirsk farm fox experiment showed that by applying strict, directional selection to a population of fur foxes (i.e., ‘wild’ animals) by always using a given generation’s tamest 10% for further breeding, a relatively fast change can be achieved towards the desirable behavioural phenotype [242]. In the case of our proposed strategies, developing amicable companion dogs at a reasonable speed could also be achieved if breeders selected only the phenotypically best specimens for further breeding. This procedure could also be aided by genetic testing if the genetic background for the desired traits were known. However, the exclusion of a larger proportion of the produced puppies from the breeding program would require administrative measures whereby those dogs with unsuitable behavioural traits could not be used later for breeding. Dog breeding is largely a leisure hobby, with a much less organized, regulated and commercialized structure than is normally the case for livestock breeding (e.g., [243]). Therefore, any of the proposed strategies for the behavioural improvement of dogs would require considerable civilian support, as well as professional guidance from the scientific community and veterinarians; predominantly the interest and enthusiasm of dog breeding organizations and the breeders themselves are imperative.
Each of the strategies would use and have consequences for mixed-breed dogs, which are not only numerous but also well-liked members of the worldwide companion dog population. The acquisition of mixed breeds can be motivated by (1) monetary reasons, (2) the wish to adopt dogs from shelters [244], and (3) by the belief that mixed breeds are healthier than purebreds [70]. There is empirical evidence that mongrels are at least as trainable and attentive as purebreds [121,245]; however, it was also shown that mixed-breed dogs show more problematic behaviours [121] and read human visual signals less effectively than purebreds do [79]. As the behavioural assessment of dogs regarding the desirable ‘amicable’ traits would be an indispensable tool in the ‘Converge’, ‘Split’ and ‘Merge’ strategies, mixed breeds could also be subjected to this procedure. The well-performing individuals could be either labelled as highly suitable candidates for the ideal urban companion role or used for the development of new dog breeds under the principles of the ‘Merge’ strategy. At the same time, as the proposed strategies are primarily aimed at the selection of improved purebred dogs (especially in the case of the ‘Converge’, ‘Diverge’ and ‘Split’ strategies), the popularizing efforts would also predominantly aim at these instead of mixed breeds. Hopefully, the pronounced need for amicable companion dogs would also have an indirect effect on stressing the importance of pre-adoption behavioural testing of mixed-breed dogs [246].
Finally, we highlight the crucial importance of the owners themselves in each of the proposed strategies for companion dog improvement. Prospective owners represent the ultimate driving force in the production and offering of dogs to the ‘market’, where breeders and shelters serve as the source. If seekers of new dogs exert their educated demand for such dogs that reliably show the traits of an amicable urban companion, the message will be clear to those on the providers’ side. Current data shows a mixed picture of dog acquisition reasons by dog owners: while the look and behaviour of the dog is important to them (e.g., [80]), just as often, they based their choice on sentimental reasons or momentary impulses (e.g., [82,133]). In the case of our proposed strategies, dog owners would need to understand that ‘amicability’ of dogs is the key component to a more enjoyable coexistence in an urban setting. With strategies ‘Converge’ and ‘Split’, we could provide a wide selection of suitable dog breeds to fulfil this need. In the framework of ‘Diverge’ and ‘Split’, dog breeds would more reliably show their original behavioural traits, thus well-educated owners could safely choose those breeds that best show the characteristics of an ideal urban companion. If the strategy ‘Merge’ were to be adopted, a whole new genre of purpose-bred companion breeds could serve the needs of knowledgeable dog owners.
Owner education programs are well known and increasingly used for various reasons. There have been initiatives for educating dog owners how to better ensure their dogs’ general health and welfare [212,247], combat canine obesity [248], or manage aggressive behaviour [249]. Lately, considerable efforts have been made to explain to fanciers of extreme brachycephalic dogs the poor welfare prospects of these breeds [250]. As the need for easy-to-live-with urban companion dogs is based on actual surveys among dog owners [64,65,185], we can be optimistic about the potential success of a campaign that would disseminate the positive features of any of the four strategies that target the development of suitable amicable companion dogs. Unlike in the case of extreme brachycephalic dogs, where the owners’ preferences for their favourite breeds have so far proven to be resistant to the grim reality of poor animal welfare (e.g., [214,250]), the need for amicable dogs in cities mirrors a public demand, which can give hope for the greater success of such initiatives.
8. Limitations and Potential Difficulties
In this paper we have proposed and assessed a few ideas that, in our opinion, could give coherent guidelines for organized breeding efforts, resulting in more suitable dogs for urban companion purposes. In some cases (‘Diverge’ and ‘Split’) these strategies could even provide working dogs that reliably show their original function-appropriate behavioural traits. However, the widespread practical success of these strategies heavily depends on many factors, and the proposed breeding strategies also could incur ethical concerns. We briefly mention the most relevant of these here.
8.1. Genetic Background of Dog Behaviour
Although there are a few promising examples of remarkably successful directional selection for sought-after behavioural traits (e.g., lessened aggression of Rottweilers in the Netherlands: [200]), to our current knowledge, most molecular genetic studies have detected a polygenic background [251,252] and widespread pleiotropy behind the behavioural phenotypes of dogs [95,151,208]. This represents serious difficulties when breeders wish to select for rather complex behavioural traits that would be preferred in an ‘ideal amicable urban dog’.
8.2. The Difficulty of Large-Scale Dog Breeding Programs
Problems arising from polygenic behavioural traits, low heritability and influence from the environment, can probably be overcome by using cutting-edge molecular genetics, large numbers of breeding animals, centralized breeding plans, strict selection criteria, and a standardized environment—as we can see in livestock science (e.g., [253]. With a few exceptions, dog breeding is a leisure activity, where most of the previously mentioned conditions are missing. There is hope that advances in canine molecular genetic research will provide guidance for more targeted selection. However, for successful, large-scale breeding programs a voluntary association of numerous devoted breeders would be needed, aided by such progressive breeding organizations (kennel clubs) who embrace the registration of the resulting dogs. The monetary background of such endeavours is also a crucial factor: many non-commercial dog breeders would need to finance the initial stage of the common breeding project; meanwhile the expected beneficial characteristics of the resulting dogs should be popularized among potential customers.
8.3. Potential Loss of Genetic Variability
Since in most dog breeds the studbooks are closed, outcrossing with breed-typical landraces, or other dog breeds is practically impossible. This, coupled with maladaptive breeding practices (e.g., using ‘popular sires’: [225]), has resulted in a significant loss of genetic diversity in many dog breeds. As some of our proposed breeding strategies necessarily include directional selection towards desired behavioural phenotypes, this may cause a further loss of genetic variability in the resulting dogs. This warrants a coordinated effort by multiple breeders who can maintain genetically separate breeding lines with the desired focus traits.
8.4. Ethical Considerations Regarding Shelter Dog Populations
The proposed breeding strategies in this paper focus on the improvement of dog breeds to provide predictably behaving, suitable urban companion dogs. The ‘Merge’ strategy even outlines the idea of creating new, amicable companion dog breeds through the use of behavioural assessment, targeted crossbreeding, and follow-up selection. However, we should not forget that around the world, hundreds of thousands of dogs are waiting for adoption at pounds, shelters and various rescue organizations (e.g., [254,255]. Although a considerable number of dogs are relinquished because of a behavioural mismatch with the owners’ expectations [40], and the lack of proper behaviour assessment of dogs in shelters can pose further problems [256], our moral obligation is to prioritize the adoption of those dogs placed with rescue organizations and shelters that could be safely and reliably reintegrated into the population of companion dogs. As behavioural testing for the preferred ‘amicable’ traits is an instrumental element of our proposed breeding strategies, these methods could also be implemented in the behavioural assessment of dogs placed in shelters, thus improving their chances of being adopted.
9. Conclusions
Our relationship with dogs is the oldest association among domesticated animals, and it has undergone constant change over the past millennia. Dogs have morphed in their form and function in parallel with the needs, preferences and possibilities of the humans living with them. However, since the advent of organized dog breeding in the 19th century, an artificially created demand for breed standardization has seemingly acted (and still tries to act) against the constant development and transformation (i.e., ‘adaptation’) of dogs according to the new requirements and needs of societies and individual dog owners. Amidst the unprecedented popularity of dog keeping in the most densely inhabited areas of our world, it is our task now to give expert guidance for the selection and nurturing of a new wave of future dog breeds. The future of dog breeds will depend on how successfully they can adapt to our urban lifestyle and the need for their original function, and it is our responsibility to shape their behavioural traits by keeping in mind the benefit of these traits for both dogs and people.
To summarize the main message and conclusion of this paper, we wish to highlight the following points:
(1) Urban dog owners would prefer ‘amicable companion dogs’, which are non-aggressive, reliable with children, have low energy levels, and can be left alone for long periods of time.
(2) Most dog breeds were originally selected to perform working tasks and therefore show several suboptimal behaviours as companion dogs in the city; meanwhile, mixed breeds have less predictable behavioural traits. However, we presume that it would be possible to genetically select for the desired ‘amicable dog’ traits in most dog breeds, aided by appropriate behavioural testing of existing and future-bred dogs as well.
(3) We propose four strategies for creating a new genre of dog breeds through purposeful breeding, with the goal of providing reliable, safe companion dogs. These dogs could hopefully satisfy owners’ needs for amicable companion animals, and they would also be better suited to life in urban settings due to their lower incidence of problematic behaviours, thereby elevating their welfare.
(4) Strategy ‘Converge’ would encourage dog breeders to exert directional selection towards the ‘amicable’ traits in most currently existing breeds. Thus, dog owners could have access to many suitable breeds with ideal behavioural traits for urban living. This strategy would be especially suitable for those breeds that already show promising companion-dog behavioural traits, and for those that show no or only a few strongly suboptimal features.
(5) Strategy ‘Diverge’ would enhance the original behavioural characteristics of the current dog breeds; thus, owners could more easily identify those breeds that are the most suitable for city living, and these dogs would reliably show the desired behaviour characteristic of their breed. This strategy would be especially advisable in the case of those dog breeds that are still commonly employed to perform their original tasks or are heavily involved in professional dog sports. Additionally, dog breeds that show mainly suboptimal behavioural traits for being an amicable urban companion would also benefit more from the ‘Diverge’ method.
(6) The strategy ‘Split’ combines ‘Converge’ and ‘Diverge’ within the breeds; thus, it helps to create a companion line and at the same time maintain a reliable working line. This strategy would be very useful for those breeds that are both wanted as companions and professional working dogs; however, the two functions require markedly different behavioural characteristics.
(7) Strategy ‘Merge’ would initiate the creation of new dog breeds that show the ideal urban companion features. For the selection, crossbreeding of current breeds, as well as mixed-breed dogs with preferred traits, could be used. This is probably the most intriguing and promising strategy, as it could perhaps be popularized most effectively. However, the creation of new companion dog breeds should be ethically justifiable due to the high number of adoptable dogs available from rescue organizations.
(8) For the success of any of these strategies, the support of many devoted breeders, breeding organizations, and expert coordinating councils would be required.
(9) As the need for easy-going, amicable urban companion dogs is based on actual demand from dog owners, we believe that the new dogs resulting from these strategies could become widely popular. However, to achieve this, whichever strategy breeders choose, widespread and well-disseminated education programs are needed for prospective dog fanciers, preferably from an early age.
Acknowledgments
The authors are grateful to Celeste R. Pongrácz, who proofread the manuscript.
Author Contributions
Conceptualization: P.P., P.D.; Investigation: P.P., P.D.; Methodology: P.P.; Project administration: P.P.; Supervision: P.P.; Visualization: P.P., P.D.; Writing—original draft: P.P., P.D.; Writing—review and editing: P.P., P.D. All authors have read and agreed to the published version of the manuscript.
Institutional Review Board Statement
Not applicable.
Informed Consent Statement
Not applicable.
Data Availability Statement
No new data were created or analyzed in this study. Data sharing is not applicable to this article.
Conflicts of Interest
The authors declare no conflicts of interest.
Funding Statement
Péter Pongrácz and Petra Dobos were supported by the Hungarian National Research, Development and Innovation Office (NKFIH, Grant # K143077), and by the Morris Animal Foundation (D25CA-039). Petra Dobos was supported by the University Research Fellowships (EKÖP-25-3-I-ELTE-147; EKÖP-26-3-I-ELTE-2026012) of the Ministry for Innovation and Technology.
Footnotes
[内容请查看原文]
References
[内容请查看原文]
Associated Data
Data Availability Statement
No new data were created or analyzed in this study. Data sharing is not applicable to this article.